An Introduction to Materials Engineering and Science for Chemical and Materials Engineers [1 ed.] 9780471436232, 0471436232

An Introduction to Materials Engineering and Science for Chemical and Materials Engineers provides a solid background in

621 21 17MB

English Pages 970 Year 2003

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

An Introduction to Materials Engineering and Science for Chemical and Materials Engineers [1 ed.]
 9780471436232, 0471436232

  • Commentary
  • 47431
  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

AN INTRODUCTION TO MATERIALS ENGINEERING AND SCIENCE

AN INTRODUCTION TO MATERIALS ENGINEERING AND SCIENCE FOR CHEMICAL AND MATERIALS ENGINEERS

Brian S. Mitchell Department of Chemical Engineering, Tulane University

A JOHN WILEY & SONS, INC., PUBLICATION

This book is printed on acid-free paper. Copyright  2004 by John Wiley & Sons, Inc., Hoboken, New Jersey. All rights reserved. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400, fax 978-750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, e-mail: [email protected]. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services please contact our Customer Care Department within the U.S. at 877-762-2974, outside the U.S. at 317-572-3993 or fax 317-572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print, however, may not be available in electronic format. Library of Congress Cataloging-in-Publication Data: Mitchell, Brian S., 1962An introduction to materials engineering and science: for chemical and materials engineers Brian S. Mitchell p. cm. Includes bibliographical references and index. ISBN 0-471-43623-2 (cloth) 1. Materials science. I. Title. TA403.M685 2003 620.1′ 1—dc21

Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1

2003053451

To my parents; whose Material was loam; Engineering was labor; Science was lore; And greatest product was love.

CONTENTS

Preface

xi

Acknowledgments

xv

1 The Structure of Materials 1.0 1.1 1.2 1.3 1.4 1.5

Introduction and Objectives Structure of Metals and Alloys Structure of Ceramics and Glasses Structure of Polymers Structure of Composites Structure of Biologics References Problems

2 Thermodynamics of Condensed Phases 2.0 2.1 2.2 2.3 2.4 2.5

Introduction and Objectives Thermodynamics of Metals and Alloys Thermodynamics of Ceramics and Glasses Thermodynamics of Polymers Thermodynamics of Composites Thermodynamics of Biologics References Problems

3 Kinetic Processes in Materials 3.0 3.1 3.2 3.3 3.4 3.5

Introduction and Objectives Kinetic Processes in Metals and Alloys Kinetic Processes in Ceramics and Glasses∗ Kinetic Processes in Polymers Kinetic Processes in Composites∗ Kinetic Processes in Biologics∗ References Problems

4 Transport Properties of Materials 4.0 Introduction and Objectives 4.1 Momentum Transport Properties of Materials∗

1 1 28 55 76 99 114 128 130 136 136 140 165 191 200 204 209 211 215 215 219 233 246 269 277 280 282 285 285 287 vii

viii

5

CONTENTS

4.2 Heat Transport Properties of Materials 4.3 Mass Transport Properties of Materials∗ References Problems

316 343 374 376

Mechanics of Materials

380

5.0 5.1 5.2 5.3 5.4 5.5

6

7

380 381 422 448 472 515 532 533

Electrical, Magnetic, and Optical Properties of Materials

537

6.1 Electrical Properties of Materials 6.2 Magnetic Properties of Materials 6.3 Optical Properties of Materials References Problems

538 600 644 677 678

Processing of Materials

681

Introduction Processing of Processing of Processing of Processing of Processing of References Problems

681 681 704 754 795 804 811 812

7.0 7.1 7.2 7.3 7.4 7.5

8

Introduction and Objectives Mechanics of Metals and Alloys Mechanics of Ceramics and Glasses Mechanics of Polymers Mechanics of Composites Mechanics of Biologics References Problems

Metals and Alloys Ceramics and Glasses Polymers Composites Biologics

Case Studies in Materials Selection

814

8.0 Introduction and Objectives 8.1 Selection of Metals for a Compressed Air Tank 8.2 Selection of Ceramic Piping for Coal Slurries in a Coal Liquefaction Plant 8.3 Selection of Polymers for Packaging 8.4 Selection of a Composite for an Automotive Drive Shaft 8.5 Selection of Materials as Tooth Coatings References Problems

814 821 827 832 835 842 848 849

CONTENTS

ix

Appendix 1:

Energy Values for Single Bonds

851

Appendix 2:

Structure of Some Common Polymers

852

Appendix 3:

Composition of Common Alloys

856

Appendix 4:

Surface and Interfacial Energies

869

Appendix 5:

Thermal Conductivities of Selected Materials

874

Appendix 6:

Diffusivities in Selected Systems

880

Appendix 7:

Mechanical Properties of Selected Materials

882

Appendix 8:

Electrical Conductivity of Selected Materials

893

Appendix 9:

Refractive Index of Selected Materials

900

Answers to Selected Problems

903

Index

907



Sections marked with an asterisk can be omitted in an introductory course.

PREFACE

This textbook is intended for use in a one- or two-semester undergraduate course in materials science that is primarily populated by chemical and materials engineering students. This is not to say that biomedical, mechanical, electrical, or civil engineering students will not be able to utilize this text, nor that the material or its presentation is unsuitable for these students. On the contrary, the breadth and depth of the material covered here is equivalent to most “traditional” metallurgy-based approaches to the subject that students in these disciplines may be more accustomed to. In fact, the treatment of biological materials on the same level as metals, ceramics, polymers, and composites may be of particular benefit to those students in the biologically related engineering disciplines. The key difference is simply the organization of the material, which is intended to benefit primarily the chemical and materials engineer. This textbook is organized on two levels: by engineering subject area and by materials class, as illustrated in the accompanying table. In terms of topic coverage, this organization is transparent: By the end of the course, the student will have covered many of the same things that would be covered utilizing a different materials science textbook. To the student, however, the organization is intended to facilitate a deeper understanding of the subject material, since it is presented in the context of courses they have already had or are currently taking—for example, thermodynamics, kinetics, transport phenomena, and unit operations. To the instructor, this organization means that, in principle, the material can be presented either in the traditional subject-oriented sequence (i.e., in rows) or in a materials-oriented sequence (i.e., in columns). The latter approach is recommended for a two-semester course, with the first two columns covered in the first semester and the final three columns covered in the second semester. The instructor should immediately recognize that the vast majority of “traditional” materials science concepts are covered in the columns on metals and ceramics, and that if the course were limited to concepts on these two materials classes only, the student would receive instruction in many of the important topics covered in a “traditional” course on materials. Similarly, many of the more advanced topics are found in the sections on polymers, composites, and biological materials and are appropriate for a senior-level, or even introductory graduate-level, course in materials with appropriate supplementation and augmentation. This textbook is further intended to provide a unique educational experience for the student. This is accomplished through the incorporation of instructional objectives, active-learning principles, design-oriented problems, and web-based information and visualization utilization. Instructional objectives are included at the beginning of each chapter to assist both the student and the instructor in determining the extent of topics and the depth of understanding required from each topic. This list should be used as a guide only: Instructors will require additional information they deem important or eliminate topics they deem inappropriate, and students will find additional topic coverage in their supplemental reading, which is encouraged through a list of references at the end xi

xii

PREFACE

Metals & Alloys

Ceramics & Glasses

Structure

Crystal structures, Point defects, Dislocations

Thermodynamics

Polymers

Composites

Biologics

Crystal structures, Defect reactions, The glassy state

Configuration, Conformation, Molecular Weight

Matrices, Reinforcements

Biochemistry, Tissue structure

Phase equilibria, Gibbs Rule Lever Rule

Ternary systems, Surface energy, Sintering

Phase separation, Polymer solutions, Polymer blends

Adhesion, Cohesion, Spreading

Cell Adhesion, Cell spreading

Kinetics

Transformations, Corrosion

Devitrification, Polymerization, Nucleation, Degradation Growth

Deposition, Infiltration

Receptors, Ligand binding

Transport Properties

Inviscid systems, Heat capacity, Diffusion

Newtonian flow, Heat capacity, Diffusion

non-Newtonian flow, Heat capacity, Diffusion

Porous Flow, Heat capacity, Diffusion

Convection, Diffusion

Mechanical Properties

Stress-strain, Elasticity, Ductility

Fatigue, Fracture, Creep

Viscoelasticity, Elastomers

Laminates

Sutures, Bone, Teeth

Electrical, Magnetic & Optical Properties

Resistivity, Magnetism, Reflectance

Dielectrics, Ferrites, Absorbance

Ion conductors, Molecular magnets, LCDs

Dielectrics, Storage media

Biosensors, MRI

Processing

Casting, Rolling, Compaction

Pressing, CVD/CVI, Sol-Gel

Extrusion, Injection molding, Blow molding

Pultrusion, RTM, CVD/CVI

Surface modification

Case Studies

Compressed air tank

Ceramic piping

Polymeric packaging

Composite drive shaft

Tooth coatings

of each chapter. Active-learning principles are exercised through the presentation of example problems in the form of Cooperative Learning Exercises. To the student, this means that they can solve problems in class and can work through specific difficulties in the presence of the instructor. Cooperative learning has been shown to increase the level of subject understanding when properly utilized.∗ No class is too large to allow students to take 5–10 minutes to solve these problems. To the instructor, the Cooperative Learning Exercises are to be used only as a starting point, and the instructor is encouraged to supplement his or her lecture with more of these problems. Particularly difficult concepts or derivations are presented in the form of Example Problems that the instructor can solve in class for the students, but the student is encouraged to solve these problems during their own group or individual study time. Design-oriented problems are offered, primarily in the Level III problems at the end of each chapter, ∗

Smith, K. Cooperative Learning and College Teaching, 3(2), 10–12 (1993).

PREFACE

xiii

that incorporate concepts from several chapters, that involve significant information retrieval or outside reading, or that require group activities. These problems may or may not have one “best” answer and are intended to promote a deeper level of understanding of the subject. Finally, there is much information on the properties of materials available on the Internet. This fact is utilized through the inclusion of appropriate web links. There are also many excellent visualization tools available on the Internet for concepts that are too difficult to comprehend in a static, two-dimensional environment, and links are provided to assist the student in their further study on these topics. Finally, the ultimate test of the success of any textbook is whether or not it stays on your bookshelf. It is hoped that the extent of physical and mechanical property data, along with the depth with which the subjects are presented, will serve the student well as they transition to the world of the practicing engineer or continue with their studies in pursuit of an advanced degree. BRIAN S. MITCHELL Tulane University

ACKNOWLEDGMENTS

The author wishes to thank the many people who have provided thoughtful input to the content and presentation of this book. In particular, the insightful criticisms and comments of Brian Grady and the anonymous reviewers are very much appreciated. Thanks also go to my students who have reviewed various iterations of this textbook, including Claudio De Castro, Shawn Haynes, Ryan Shexsnaydre, and Amanda Moster, as well as Dennis Johnson, Eric Hampsey, and Tom Fan. The support of my colleagues during the writing of this book, along with the support of the departmental staff, are gratefully acknowledged. Finally, the moral support of Bonnie, Britt, Rory, and Chelsie is what ultimately has led to the completion of this textbook—thank you. BRIAN S. MITCHELL Tulane University

xv

CHAPTER 1

The Structure of Materials

1.0

INTRODUCTION AND OBJECTIVES

A wealth of information can be obtained by looking at the structure of a material. Though there are many levels of structure (e.g., atomic vs. macroscopic), many physical properties of a material can be related directly to the arrangement and types of bonds that make up that material. We will begin by reviewing some general chemical principles that will aid us in our description of material structure. Such topics as periodic structure, types of bonding, and potential energy diagrams will be reviewed. We will then use this information to look at the specific materials categories in more detail: metals, ceramics, polymers, composites, and biological materials (biologics). There will be topics that are specific to each material class, and there will also be some that are common to all types of materials. In subsequent chapters, we will explore not only how the building blocks of a material can significantly impact the properties a material possesses, but also how the material interacts with its environment and other materials surrounding it. By the end of this chapter you should be able to: ž ž ž ž ž ž ž ž ž ž ž ž

Identify trends in the periodic table for IE, EA, electronegativity, and atomic/ionic radii. Identify the type of bonding in a compound. Utilize the concepts of molecular orbital and hybridization theories to explain multiple bonds, bond angle, diamagnetism, and paramagnetism. Identify the seven crystal systems and 14 Bravais lattices. Calculate the volume of a unit cell from the lattice translation vectors. Calculate atomic density along directions, planes, and volumes in a unit cell. Calculate the density of a compound from its crystal structure and atomic mass. Locate and identify the interstitial sites in a crystal structure. Assign coordinates to a location, indices to a direction, and Miller indices to a plane in a unit cell. Use Bragg’s Law to convert between diffraction angle and interplanar spacing. Read and interpret a simple X-ray diffraction pattern. Identify types of point and line defects in solids.

An Introduction to Materials Engineering and Science: For Chemical and Materials Engineers, by Brian S. Mitchell ISBN 0-471-43623-2 Copyright  2004 John Wiley & Sons, Inc.

1

2

THE STRUCTURE OF MATERIALS

ž ž ž ž ž ž ž ž ž ž ž ž ž

Calculate the concentration of point defects in solids. Draw a Burger’s circuit and identify the direction of dislocation propagation. Use Pauling’s rules to determine the stability of a compound. Predict the structure of a silicate from the Si/O ratio. Apply Zachariasen’s rules to determine the glass forming ability of an oxide. Write balanced defect reaction equations using Kroger–Vink notation. Classify polymers according to structure or formability. Calculate the first three moments of a polymer molecular weight distribution. Apply principles of glass transition and polymer crystallinity to polymer classification. Identify nematic, smectic, and cholesteric structures in liquid crystalline polymers. Identify the components in a composite material. Approximate physical properties of a composite material based on component properties. Be conversant in terms that relate to the structure of biological materials, such as fibronectin and integrins.

1.0.1

The Elements

Elements are materials, too. Oftentimes, this fact is overlooked. Think about all the materials from our daily lives that are elements: gold and silver for our jewelry; aluminum for our soda cans; copper for our plumbing; carbon, both as a luminescent diamond and as a mundane pencil lead; mercury for our thermometers; and tungsten for our light bulb filaments. Most of these elements, however, are relatively scarce in the grand scheme of things. A table of the relative abundance of elements (Table 1.1) shows that most of our universe is made up of hydrogen and helium. A little closer to home, things are much different. A similar table of relative abundance (Table 1.2) shows that helium on earth is relatively scarce, while oxygen dominates the crust of our planet. Just think of how much molecular oxygen, water, and aluminosilicate rocks are contained in the earth’s crust. But those are molecules—we are concentrating on atoms for the moment. Still, elements are of vital importance on earth, and the ones we use most often are primarily in the solid form. Recall from your introductory chemistry course that the elements can be systematically arranged in a periodic table according to their electronic structure (see Table 1.3∗ ). An overall look at the periodic table (Figure 1.1) shows that many elements are solids (white boxes) at room temperature. The fact that many of these elements remain solid well above ambient temperatures is important. As we heat to 1000◦ C, note that many of the IIIA–VA elements have melted (light shaded); also note how many of the alkali metals (IA) have vaporized (dark shaded), but how most of the transition elements are still in solid form. At 2000◦ C, the alkali earths are molten, and many of the transition elements have begun to melt, too. Note that the highest melting point element is carbon (Figure 1.1d). Keep in mind that this is in an inert atmosphere. What should happen to ∗

Note that the Lanthanide (atomic numbers 58–71) and Actinide (90–103) series elements, as well as the synthetic elements of atomic number greater than 87, are omitted from all the periodic tables in this text. With the possible exception of nuclear fuels such as uranium and plutonium, these elements are of little general engineering interest.

INTRODUCTION AND OBJECTIVES

Table 1.1 Universe

3

Relative Abundance of Elements in the

Relative Abundance (Si = 1)

Element Hydrogen (H) Helium (He) Oxygen (O) Nitrogen (N) Carbon (C) Iron (Fe) Silicon (Si) Magnesium (Mg) Sulfur (S) Nickel (Ni) Aluminum (Al) Calcium (Ca) Sodium (Na) Chlorine (Cl)

12,000 2,800 16 8 3 2.6 1 0.89 0.33 0.21 0.09 0.07 0.045 0.025

Table 1.2 Relative Abundance of Selected Elements in the Earth’s Crust Element Oxygen (O) Silicon (Si) Aluminum (Al) Iron (Fe) Calcium (Ca) Sodium (Na) Potassium (K) Magnesium (Mg) Titanium (Ti) Hydrogen (H) Phosphorus (P) Manganese (Mn) Sulfur (S) Carbon (C) Chlorine (Cl)

Relative Abundance (ppm) 466,000 277,200 81,300 50,000 36,300 28,300 25,900 20,900 4,400 1,400 1,180 1,000 520 320 314

Element Fluorine (F) Strontium (Sr) Barium (Ba) Zirconium (Zr) Chromium (Cr) Vanadium (V) Zinc (Zn) Nickel (Ni) Molybdenum (Mo) Uranium (U) Mercury (Hg) Silver (Ag) Platinum (Pt) Gold (Au) Helium (He)

Relative Abundance (ppm) 300 300 250 220 200 150 132 80 15 4 0.5 0.1 0.005 0.005 0.003

this element in the presence of oxygen? Such elements as tungsten, platinum, molybdenum, and tantalum have exceptional high-temperature properties. Later on we will investigate why this is so. In addition, many elements are, in and of themselves, materials of construction. Aluminum and copper are just a few examples of elements that are used extensively for fabricating mechanical parts. Elements have special electrical characteristics, too. Silver and gold are used not just for jewelry, but also for a wide variety of electrical components. We will visit all of these topics in the course of this textbook.

4

THE STRUCTURE OF MATERIALS

1.0.2

Trends in the Periodic Table

A closer look at the periodic table points out some interesting trends. These trends not only help us predict how one element might perform relative to another, but also give us some insight into the important properties of atoms and ions that determine their performance. For example, examination of the melting points of the elements in Table 1.3 shows that there is a general trend to decrease melting point as we go down

1 H 3 4 Li Be 11 12 Na Mg

Temperature: 290 °K 16 °C 62 °F

24 25 26 Cr Mn Fe

5 B 13 Al 27 Co

31 32 33 Ga Ge As

34 Se

35 Br

36 Kr

47 Ag 79 Au

48 Cd 80 Hg

49 ln 81 TI

50 51 Sn Sb 82 83 Pb Bi

52 Te 84 Po

53 I 85 At

54 Xe 86 Rn

7 N 15 P

8 O 16 S

2 He 9 10 F Ne 17 18 CI Ar

21 Sc

22 Ti

23 V

37 Rb 55 Cs

38 Sr 56 Ba

39 Y 57 La

40 Zr 72 Hf

41 42 43 44 45 Nb Mo Tc Ru Rh 73 74 75 76 77 Ta W Re Os Ir

46 Pd 78 Pt

87 Fr

88 Ra

89 Ac

59 Pr 91 Pa

60 61 62 63 Nd Pm Sm Eu 92 93 94 95 U Np Pu Am

65 66 Tb Dy 97 98 Bk Cf

64 Gd 96 Cm

8 O 16 S

30 Zn

20 Ca

58 Ce 90 Th

7 N 15 P

28 29 Ni Cu

19 K

Legend Solid Liquid Gas Not Available

6 C 14 Si

2 He 9 10 F Ne 17 18 CI Ar

67 68 69 Ho Er Tm 99 100 101 Es Fm Md

70 Yb 102 No

71 Lu 103 Lr

(a)

1 H 3 4 Li Be 11 12 Na Mg

Temperature: 1280 °K 1006 °C 1844 °F

5 B 13 Al 28 29 Ni Cu

30 Zn

31 32 33 Ga Ge As

34 Se

35 Br

36 Kr

47 Ag 79 Au

48 Cd 80 Hg

49 ln 81 TI

50 51 Sn Sb 82 83 Pb Bi

52 Te 84 Po

53 I 85 At

54 Xe 86 Rn

19 K

20 Ca

21 Sc

22 Ti

23 V

37 Rb 55 Cs

38 Sr 56 Ba

39 Y 57 La

40 Zr 72 Hf

41 42 43 44 45 Nb Mo Tc Ru Rh 73 74 75 76 77 Ta W Re Os Ir

46 Pd 78 Pt

87 Fr

88 Ra

89 Ac

59 Pr 91 Pa

60 61 62 63 Nd Pm Sm Eu 92 93 94 95 U Np Pu Am

65 66 Tb Dy 97 98 Bk Cf

Legend Solid Liquid Gas Not Available

58 Ce 90 Th

24 25 26 Cr Mn Fe

27 Co

64 Gd 96 Cm

6 C 14 Si

67 68 69 Ho Er Tm 99 100 101 Es Fm Md

70 Yb 102 No

71 Lu 103 Lr

(b)

Figure 1.1 The periodic table of the elements at (a) room temperature, (b) 1000◦ C, (c) 2000◦ C, and (d) 3500◦ C.

INTRODUCTION AND OBJECTIVES

1 H 3 4 Li Be 11 12 Na Mg

Temperature: 2280 °K 2006 °C 3644 °F

19 K

20 Ca

21 Sc

22 Ti

23 V

37 Rb 55 Cs

38 Sr 56 Ba

39 Y 57 La

40 Zr 72 Hf

87 Fr

88 Ra

89 Ac

59 Pr 91 Pa

Legend Solid Liquid Gas Not Available

58 Ce 90 Th

5 B 13 Al

6 C 14 Si

7 N 15 P

8 O 16 S

2 He 9 10 F Ne 17 18 CI Ar

28 29 Ni Cu

30 Zn

31 32 33 Ga Ge As

34 Se

35 Br

36 Kr

41 42 43 44 45 Nb Mo Tc Ru Rh 73 74 75 76 77 Ta W Re Os Ir

46 Pd 78 Pt

48 Cd 80 Hg

49 ln 81 TI

52 Te 84 Po

53 I 85 At

54 Xe 86 Rn

60 61 62 63 Nd Pm Sm Eu 92 93 94 95 U Np Pu Am

65 66 Tb Dy 97 98 Bk Cf

7 N 15 P

8 O 16 S

2 He 9 10 F Ne 17 18 CI Ar

24 25 26 Cr Mn Fe

27 Co

64 Gd 96 Cm

47 Ag 79 Au

67 68 69 Ho Er Tm 99 100 101 Es Fm Md

50 51 Sn Sb 82 83 Pb Bi

70 Yb 102 No

5

71 Lu 103 Lr

(c)

1 H 3 4 Li Be 11 12 Na Mg

Temperature: 3780 °K 3506 °C 6344 °F

19 K

20 Ca

21 Sc

22 Ti

23 V

37 Rb 55 Cs

38 Sr 56 Ba

39 Y 57 La

40 Zr 72 Hf

87 Fr

88 Ra

89 Ac

59 Pr 91 Pa

Legend Solid Liquid Gas Not Available

58 Ce 90 Th

5 B 13 Al

6 C 14 Si

28 29 Ni Cu

30 Zn

31 32 33 Ga Ge As

34 Se

35 Br

36 Kr

41 42 43 44 45 Nb Mo Tc Ru Rh 73 74 75 76 77 Ta W Re Os Ir

46 Pd 78 Pt

48 Cd 80 Hg

49 ln 81 TI

52 Te 84 Po

53 I 85 At

54 Xe 86 Rn

60 61 62 63 Nd Pm Sm Eu 92 93 94 95 U Np Pu Am

65 66 Tb Dy 97 98 Bk Cf

24 25 26 Cr Mn Fe

27 Co

64 Gd 96 Cm

47 Ag 79 Au

67 68 69 Ho Er Tm 99 100 101 Es Fm Md

50 51 Sn Sb 82 83 Pb Bi

70 Yb 102 No

71 Lu 103 Lr

(d)

Figure 1.1 (continued ).

a column for the alkali metals and alkali earth elements (columns IA and IIA), but that the column trend for the transition metals appears to be different. There are some trends that are more uniform, however, and are related to the electronic structure of the element. 1.0.2.1 First Ionization Energy (IE). The first ionization energy, abbreviated IE, is sometimes referred to as the “ionization potential.” It is the energy required to remove

12 Mg Ne3s 2 24.30 649

20 Ca Ar4s 2 40.08 840

38 Sr Kr5s 2 87.62 769

56 Ba Xe6s 2 137.3 729

11 Na Ne3s 1 22.99 97.8

19 K Ar4s 1 39.10 63.2

37 Rb Kr5s 1 85.47 39.5

55 Cs Xe6s 1 132.9 28.4

es aw mp

es aw mp

es aw mp

es aw mp

es aw mp

57 La Ba5d 1 138.9 921

39 Y Sr4d 1 88.91 1528

21 Sc Ca3d 1 44.96 1541

72 Hf Ba5d 2 178.5 2231

40 Zr Sr4d 2 91.22 1865

22 Ti Ca3d 2 47.9 1672

73 Ta Ba5d 3 180.9 3020

41 Nb Rb4d 4 92.91 2471

23 V Ca3d 3 50.94 1929

74 W Ba5d 4 183.9 3387

42 Mo Rb4d 5 95.94 2623

24 Cr K3d 5 51.9 1863

75 Re Ba5d 5 186.2 3186

43 Tc Rb4d 6 98.91 2204

25 Mn Ca3d 5 54.93 1246

76 Os Ba5d 6 190.2 3033

44 Ru Rb4d 7 101.7 2254

26 Fe Ca3d 6 55.85 1538

Source: Ralls, Courtney, Wulff, Introduction to Materials Science and Engineering, Wiley, 1976.

4 Be He2s 2 9.012 1289

3 Li He2s 1 6.94 180.5

es aw mp

77 Ir Xe5d 9 192.2 2447

45 Rh Rb4d 8 102.9 1963

27 Co Ca3d 7 58.93 1494

78 Pt Cs5d 9 195.1 1772

46 Pd Kr4d 10 106.4 1554

28 Ni Ca3d 8 58.71 1455

es = electronic structure aw = atomic weight (average including isotopes) mp = melting point,◦ C (sublimation temperatures enclosed in parentheses).

Electronic Structure and Melting Points of the Elements

1 H 1s 1 1.008 −259

Table 1.3

79 Au Cs5d 10 196.97 1064.4

47 Ag Rb4d 10 107.87 961.9

29 Cu K3d 10 63.55 1084.5

80 Hg Ba5d 10 200.6 −38.9

48 Cd Sr4d 10 112.4 321.1

30 Zn Ca3d 10 65.37 419.6

81 Tl Ba6p 1 204.4 304

49 In Sr5p 1 114.8 156.6

31 Ga Ca4p 1 69.72 29.8

13 Al Mg3p 1 26.98 660.4

5 B Be2p 1 10.81 ∼2103

82 Pb Ba6p 2 207.2 327.5

50 Sn Sr5p 2 118.7 232.0

32 Ge Ca4p 2 72.59 938.3

14 Si Mg3p 2 28.09 1414

6 C Be2p 2 12.01 (3836)

83 Bi Ba6p 3 208.9 271.4

51 Sb Sr5p 3 121.8 630.7

33 As Ca4p 3 74.92 (603)

15 P Mg3p 3 30.974 44.1

7 N Be2p 3 14.006 −210.0

84 Po Ba6p 4 210 254

52 Te Sr5p 4 127.6 449.6

34 Se Ca4p 4 78.96 221

16 S Mg3p 4 32.06 112.8

8 O Be2p 4 15.999 −218.8

85 At Ba6p 5 210 —

53 I Sr5p 5 126.9 113.6

35 Br Ca4p 5 79.90 −7.2

17 Cl Mg3p 5 35.45 −101.0

9 F Be2p 5 18.998 −219.7

86 Rn Ba6p 6 222 −71

54 Xe Sr5p 6 131.3 −112

36 Kr Ca4p 6 83.80 −157

18 Ar Mg3p 6 39.95 −189

10 Ne Be2p 6 20.18 −249

2 He 1s 2 4.003 —

6 THE STRUCTURE OF MATERIALS

7

INTRODUCTION AND OBJECTIVES

the most weakly bound (usually outermost) electron from an isolated gaseous atom positive ion (g) + e−

atom (g) + IE

(1.1)

and can be calculated using the energy of the outermost electron as given by the Bohr model and Schr¨odinger’s equation (in eV): IE =

13.6Z 2 n2

(1.2)

where Z is the effective nuclear charge and n is the principal quantum number. As shown in Figure 1.2a, the general trend in the periodic table is for the ionization energy to increase from bottom to top and from left to right (why?). A quantity related to the IE is the work function. The work function is the energy necessary to remove an electron from the metal surface in thermoelectric or photoelectric emission. We will describe this in more detail in conjunction with electronic properties of materials in Chapter 6.

(a)

(b)

(c)

(d)

Figure 1.2 Some important trends in the periodic table for (a) ionization energy, (b) electron affinity, (c) atomic and ionic radii, and (d) electronegativity. Increasing values are in the direction of the arrow.

8

THE STRUCTURE OF MATERIALS

1.0.2.2 Electron Affinity (EA). Electron affinity is the reverse process to the ionization energy; it is the energy change (often expressed in eV) associated with an isolated gaseous atom accepting one electron:

atom (g) + e−

negative ion (g)

(1.3)

Unlike the ionization energy, however, EA can have either a negative or positive value, so it is not included in Eq. (1.3). The EA is positive if energy is released upon formation of the negative ion. If energy is required, EA is negative. The general trend in the periodic table is again toward an increase in EA as we go from the bottom to top, and left to right (Figure 1.2b), though this trend is much less uniform than for the IE. 1.0.2.3 Atomic and Ionic Radii. In general, positive ions are smaller than neutral atoms, while negative ions are larger (why?). The trend in ionic and atomic radii is opposite to that of IE and EA (Figure 1.2c). In general, there is an increase in radius from top to bottom, right to left. In this case, the effective nuclear charge increases from left to right, the inner electrons cannot shield as effectively, and the outer electrons are drawn close to the nucleus, reducing the atomic radius. Note that the radii are only approximations because the orbitals, in theory, extend to infinity. 1.0.2.4 Electronegativity. The ionization energy and electron affinity are characteristics of isolated atoms; they say very little about how two atoms will interact with each other. It would be nice to have an independent measure of the attraction an atom has for electrons in a bond formed with another atom. Electronegativity is such a quantity. It is represented by the lowercase Greek letter “chi,” χ. Values can be calculated using one of several methods discussed below. Values of χ are always relative to one another for a given method of calculation, and values from one method should not be used with values from another method. Based upon a scale developed by Mulliken, electronegativity is the average of the ionization energy and the electron affinity:

χ=

IE + EA 2

(1.4)

There are other types of electronegativity scales as well, the most widely utilized of which is the one from the developer of the electronegativity concept, Linus Pauling: χ=

0.31(n + 1 ± c) + 0.5 r

(1.5)

where n is the number of valence electrons, c is any formal valence charge on the atom and the sign corresponding to it, and r is the covalent radius. Typical electronegativity values, along with values of IE and EA, are listed in Table 1.4. We will use the concept of electronegativity to discuss chemical bonding. 1.0.3

Types of Bonds

Electronegativity is a very useful quantity to help categorize bonds, because it provides a measure of the excess binding energy between atoms A and B, A−B (in kJ/mol): A−B = 96.5(χA − χB )2

(1.6)

9

73 Ta 577 — 1.5

74 W 770 — 2.36

75 Re 761 — 1.9

76 Os 841 — 2.2

44 Ru 724 — 2.28 77 Ir 887 — 2.2

45 Rh 745 — 2.2

Ionization energy (IE ) and Electron affinities (EA) are expressed as kilojoules per mole. 1 eV = 96,490 J/mol. Source: R. E. Dickerson, H. B. Gray, and G. P. Haight, Chemical Principles, 3rd ed., Pearson Education, Inc., 1979.

a

72 Hf 531 — 1.3

43 Tc 699 — 1.9 78 Pt 866 — 2.28

46 Pd 803 — 2.20

28 Ni 736 — 1.91

79 Au 891 — 2.54

47 Ag 732 — 1.93

29 Cu 745 — 1.90

80 Hg 1008 — 2.00

48 Cd 866 — 1.69

30 Zn 904 — 1.65

IE EA χ

57 La 540 — 1.10

42 Mo 695 — 2.16

27 Co 757 — 1.88

56 Ba 502 — 0.89

41 Nb 653 — 1.6

26 Fe 761 — 1.8

55 Cs 377 — 0.79

40 Zr 669 — 1.33

25 Mn 715 — 1.55

IE EA χ

39 Y 636 — 1.22

24 Cr 653 — 1.66

38 Sr 548 — 0.95

23 V 653 — 1.63

81 Tl 590 — 2.04

49 In 556 — 1.78

31 Ga 577 — 1.81

82 Pb 715 — 2.33

50 Sn 707 — 1.96

32 Ge 782 — 2.01

83 Bi 774 — 2.02

51 Sb 833 — 2.05

33 As 966 — 2.18

84 Po — — 2.0

52 Te 870 — 2.1

34 Se 941 — 2.55

85 At — — 2.2

53 I 1008 304.2 2.66

35 Br 1142 333.0 2.96

86 Rn 1038 — —

54 Xe 1172 — —

36 Kr 1351 — —

18 Ar 2372 −60.2 —

37 Rb 402 — 0.82

22 Ti 661 — 1.54

17 Cl 1255 356.1 3.16

IE EA χ

21 Sc 632 — 1.36

16 S 1000 199.6 2.58

20 Ca 590 — 1.00

15 P 1063 75.3 2.19

19 K 1310 67.4 2.20

14 Si 787 138.1 1.90

IE EA χ

13 Al 577 50.2 1.61

12 Mg 736 0 1.31

9 F 1682 349.4 3.98

11 Na 498 117.2 0.93

8 O 1314 141.8 3.44

IE EA χ

7 N 1406 4.6 3.04

10 Ne 2080 −54.8 —

6 C 1088 119.7 2.55

4 Be 900 −18.4 1.57

3 Li 519 77.0 0.98

IE EA χ 5 B 799 31.8 2.04

2 He 2372 −60.2 —

1 H 1310 67.4 2.20

Table 1.4 Ionization Energies, Electron Affinities, and Electronegativities of the Elementsa

10

THE STRUCTURE OF MATERIALS

The excess binding energy, in turn, is related to a measurable quantity, namely the bond dissociation energy between two atoms, DE ij : A−B = DE AB − [(DE AA )(DE BB )]1/2

(1.7)

The bond dissociation energy is the energy required to separate two bonded atoms (see Appendix 1 for typical values). The greater the electronegativity difference, the greater the excess binding energy. These quantities give us a method of characterizing bond types. More importantly, they relate to important physical properties, such as melting point (see Table 1.5). First, let us review the bond types and characteristics, then describe each in more detail. 1.0.3.1 Primary Bonds. Primary bonds, also known as “strong bonds,” are created when there is direct interaction of electrons between two or more atoms, either through transfer or as a result of sharing. The more electrons per atom that take place in this process, the higher the bond “order” (e.g., single, double, or triple bond) and the stronger the connection between atoms. There are four general categories of primary bonds: ionic, covalent, polar covalent, and metallic. An ionic bond, also called a heteropolar

Table 1.5

Examples of Substances with Different Types of Interatomic Bonding

Type of Bond

Substance

Ionic

CaCl NaCl LiF CuF2 Al2 O3

Covalent

Ge GaAs Si SiC Diamond

Metallic

Na Al Cu Fe W

van der Waals

Ne Ar CH4 Kr Cl2

Hydrogen bonding HF H2 O

Bond Energy, Melting Point, kJ/mol (◦ C)

Characteristics

651 768 1008 2591 15,192

646 801 870 1360 3500

Low electrical conductivity, transparent, brittle, high melting point

315 ∼315 353 1188 714

958 1238 1420 2600 3550

Low electrical conductivity, very hard, very high melting point

97.5 660 1083 1535 3370

High electrical and thermal conductivity, easily deformable, opaque

−248.7 −189.4 −184 −157 −103

Weak binding, low melting and boiling points, very compressible

109 311 340 407 844 2.5 7.6 10 12 31 29 50

−92 0

Higher melting point than van der Waals bonding, tendency to form groups of many molecules

INTRODUCTION AND OBJECTIVES

11

bond, results when electrons are transferred from the more electropositive atom to the more electronegative atom, as in sodium chloride, NaCl. Ionic bonds usually result when the electronegativity difference between two atoms in a diatomic molecule is greater than about 2.0. Because of the large discrepancy in electronegativities, one atom will generally gain an electron, while the other atom in a diatomic molecule will lose an electron. Both atoms tend to be “satisfied” with this arrangement because they oftentimes end up with noble gas electron configurations—that is, full electronic orbitals. The classic example of an ionic bond is NaCl, but CaF2 and MgO are also examples of molecules in which ionic bonding dominates. A covalent bond, or homopolar bond, arises when electrons are shared between two atoms (e.g., H–H). This means that a binding electron in a covalent diatomic molecule such as H2 has equal likelihood of being found around either hydrogen atom. Covalent bonds are typically found in homonuclear diatomics such as O2 and N2 , though the atoms need not be the same to have similar electronegativities. Electronegativity differences of less than about 0.4 characterize covalent bonds. For two atoms with an electronegativity difference of between 0.4 and 2.0, a polar covalent bond is formed—one that is neither truly ionic nor totally covalent. An example of a polar covalent bond can be found in the molecule hydrogen fluoride, HF. Though there is significant sharing of the electrons, some charge distribution exists that results in a polar or partial ionic character to the bond. The percent ionic character of the bond can again be related to the electronegativities of the individual atoms: % ionic character = 100{1 − exp[−0.25(χA − χB )2 ]}

(1.8)

Example Problem 1.1

What is the percent ionic character of H–F? Answer: According to Table 1.3, the electronegativity of hydrogen is 2.20 and that of fluorine 3.98. Putting these values into Eq. (1.8) gives % ionic character of H–F = 100[1 − exp{−0.25(2.20 − 3.98)2 }] = 55%

The larger the electronegativity difference, the more ionic character the bond has. Of course, if the electronegativity difference is greater than about 2.0, we know that an ionic bond should result. Finally, a special type of primary bond known as a metallic bond is found in an assembly of homonuclear atoms, such as copper or sodium. Here the bonding electrons become “decentralized” and are shared by the core of positive nuclei. Metallic bonds occur when elements of low electronegativity (usually found in the lower left region of the periodic table) bond with each other to form a class of materials we call metals. Metals tend to have common characteristics such as ductility, luster, and high thermal and electrical conductivity. All of these characteristics can to some degree be accounted for by the nature of the metallic bond. The model of a metallic bond, first proposed by Lorentz, consists of an assembly of positively charged ion cores surrounded by free electrons or an “electron gas.” We will see later on, when we

12

THE STRUCTURE OF MATERIALS

describe intermolecular forces and bonding, that the electron cloud does indeed have “structure” in the quantum mechanical sense, which accounts nicely for the observed electrical properties of these materials. 1.0.3.2 Secondary Bonds. Secondary bonds, or weak bonds, occur due to indirect interaction of electrons in adjacent atoms or molecules. There are three main types of secondary bonding: hydrogen bonding, dipole–dipole interactions, and van der Waals forces. The latter, named after the famous Dutch physicist who first described them, arise due to momentary electric dipoles (regions of positive and negative charge) that can occur in all atoms and molecules due to statistical variations in the charge density. These intermolecular forces are common, but very weak, and are found in inert gases where other types of bonding do not exist. Hydrogen bonding is the attraction between hydrogen in a highly polar molecule and the electronegative atom in another polar molecule. In the water molecule, oxygen draws much of the electron density around it, creating positively charged centers at the two hydrogen atoms. These positively charged hydrogen atoms can interact with the negative center around the oxygen in adjacent water molecules. Although this type of

HISTORICAL HIGHLIGHT

Dutch physicist Johannes Diderik van der Waals was born on November 23, 1837 in Leiden, the Netherlands. He was the eldest son of eight children. Initially, van der Waals was an elementary school teacher during the years 1856–1861. He continued studying to become headmaster and attended lectures on mathematics, physics, and astronomy at Leiden University. From 1866 onwards he taught physics and mathematics at a secondary school in The Hague. After a revision of the law, knowledge of Latin and Greek was no longer a prerequisite for an academic graduation, and in 1873 J. D. van der Waals graduated on the thesis: “Over de continu¨iteit van de gas—envloeistoftoestand” (“About the continuity of gaseous and liquid states”). In this thesis he published the well-known law: 

P+

a  (V − b) = RT V2

This revision to the ideal gas law accounted for the specific volume of gas molecules and assumed a force between these molecules which are now known as “van der Waals forces.” With this law, the existence of

condensation and the critical temperature of gases could be predicted. In 1877 J. D. van der Waals became the first professor of physics at the University “Illustre” in Amsterdam. In 1880 he formulated his “law of corresponding states,” in 1893 he devised a theory for capillary phenomena, and in 1891 he introduced his theory for the behavior of two mixtures of two materials. It was not possible to experimentally show the de-mixing of two gases into two separate gases under certain circumstances as predicted by this theory until 1941. From 1875 to 1895 J.D. van der Waals was a member of the Dutch Royal Academy of Science. In 1908, at the age of 71, J. D. van der Waals resigned as a professor. During his life J. D. van der Waals was honored many times. He was one of only 12 foreign members of the “Academie des Sciences” in Paris. In 1910 he received the Nobel prize for Physics for the incredible work he had done on the equations of state for gases and fluids—only the fifth Dutch physicist to receive this honor. J. D. van der Waals died on March 8, 1923 at the age of 85. Source: www.vdwaals.nl

INTRODUCTION AND OBJECTIVES

13

bonding is of the same order of magnitude in strength as van der Waals bonding, it can have a profound influence on the properties of a material, such as boiling and melting points. In addition to having important chemical and physical implications, hydrogen bonding plays an important role in many biological and environmental phenomena. It is responsible for causing ice to be less dense than water (how many other substances do you know that are less dense in the solid state than in the liquid state?), an occurrence that allows fish to survive at the bottom of frozen lakes. Finally, some molecules possess permanent charge separations, or dipoles, such as are found in water. The general case for the interaction of any positive dipole with a negative dipole is called dipole–dipole interaction. Hydrogen bonding can be thought of as a specific type of dipole–dipole interaction. A dipolar molecule like ammonia, NH3 , is able to dissolve other polar molecules, like water, due to dipole–dipole interactions. In the case of NaCl in water, the dipole–dipole interactions are so strong as to break the intermolecular forces within the molecular solid. Now that the types of bonds have been reviewed, we will concentrate on the primary bond because it correlates more directly with physical properties in solids than do secondary bonds. Be aware that the secondary forces exist, though, and that they play a larger role in liquids and gases than in solids. 1.0.4

Intermolecular Forces and Bonding

We have described the different types of primary bonds, but how do these bonds form in the first place? What is it that causes a sodium ion and a chloride ion to form a compound, and what is it that prevents the nuclei from fusing together to form one element? These questions all lead us to the topics of intermolecular forces and bond formation. We know that atoms approach each other only to a certain distance, and then, if they form a compound, they will maintain some equilibrium separation distance known as the bond length. Hence, we expect that there is some attractive energy that brings them together, as well as some repulsive energy that keeps the atoms a certain distance apart. Also known as chemical affinity, the attractive energy between atoms is what causes them to approach each other. This attraction is due to the electrostatic force between the nucleus and electron clouds of the separate atoms. It should make sense to you that the attractive energy (UA ) is inversely proportional to the separation distance, r; that is, the further the atoms are apart, the weaker the attraction: UA = −

a rm

(1.9)

where a is a constant that we will describe in more detail in a moment, and m is a constant with a value of 1 for ions and 6 for molecules. Notice that there is a negative sign in Eq. (1.9). By convention, we will refer to the attractive energy as a “negative energy.” Once the atoms begin to approach each other, they can only come so close together due to the impenetrability of matter. The result is a repulsive energy, which we assign a positive value, again, by convention. The primary constituents of this repulsive energy are nucleus–nucleus and electron–electron repulsions. As with the attractive energy, the repulsive energy is inversely proportional to the separation distance; the closer the

14

THE STRUCTURE OF MATERIALS

Table 1.6 Values of the Repulsion Exponent Noble Gas Ion Core

Outer Core Configuration

n

He Ne Ar Kr Xe

1s 2 2s 2 2p 6 3s 2 3p 6 3d 10 4s 2 4p 6 4d 10 5s 2 5p 6

5 7 9 10 12

atoms are, the more they repel each other: UR =

b rn

(1.10)

where b and n are constants. The value of n, called the repulsion exponent, depends on the outer core configuration of the atom. Values of the repulsion exponent are given in Table 1.6. The total, or potential energy of the system is then the sum of the attractive and repulsive components: b −a U = UA + UR = m + n (1.11) r r The result is the potential energy well (see Figure 1.3). The well-known Lennard-Jones potential b a (1.12) U = − 6 + 12 r r is a common potential energy function used in a number of models, including collision theory for kinetics. It is simply a special case of Eq. (1.11) with n = 12 (Xe configuration) and m = 6 (molecules). It is oftentimes useful to know the forces involved in bonding, as well as the energy. Recall that energy and force, F , are related by F =−

dU dr

(1.13)

We will see later on that we can use this expression to convert between force and energy for specific types of atoms and molecules (specific values of n and m). For now, this expression helps us find the equilibrium bond distance, r0 , which occurs when forces are equal (the sum of attractive and repulsive forces is zero) or at minimum potential energy (take the derivative and set it equal to zero): F =−

dU = 0, dr

at r = r0

(1.14)

This is not the same thing as the maximum attractive force, which we get by maximizing F : dF d 2U Fmax = (1.15) =− 2 =0 dr dr

INTRODUCTION AND OBJECTIVES

15

Repulsion energy, bn r s 0 r =s

rx

Interatomic or intermolecular distance in Å r The net potential energy, U = −

a b + rm rn

Attraction energy, − am r

Attraction

Potential energy, U

Repulsion

+

r0 Um −

(a)

Force, F

Repulsion

+

Repulsion force

r0 rx

Interatomic or intermolecular distance in Å r

Attraction

0

Attraction force

Fmax −

(b)

Figure 1.3 The interatomic (a) potential energy and (b) force diagrams. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. by Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

The forces are equal when the potential energy is a minimum and the separation distance is at the bond length, r0 . Differentiation of Eq. (1.11) and solving for r0 in terms of a, b, n, and m gives

r0 =



 1 nb n − m ma

(1.16)

16

THE STRUCTURE OF MATERIALS

The potential energy well concept is an important one, not only for the calculation of binding energies, as we will see in a moment, but for a number of important physical properties of materials. How tightly atoms are bound together in a compound has a direct impact on such properties as melting point, elastic modulus and thermal expansion coefficient. Figure 1.4 shows a qualitative comparison of a material that has a deep and narrow potential energy wells versus one in which the potential energy well is wide and shallow. The deeper well represents stronger interatomic attraction; hence it is more difficult to melt these substances, which have correspondingly large elastic moduli and low thermal expansion coefficients.

Cooperative Learning Exercise 1.1

Work with a neighbor. Consider the Lennard-Jones potential, as given by Eq. (1.12), for which m = 6 and n = 12. You wish to determine the separation distance, r, at which the maximum force occurs, Fmax , in terms of the equilibrium bond distance, r0 . Person 1: Use Eq. (1.16) with the values of m and n of the Lennard-Jones potential to solve for the constant a in terms of b and the equilibrium bond distance, r0 . Now perform the determination of Fmax as given by Eq. (1.15): substitute this value of a back into Eq. (1.12), differentiate it twice with respect to r (remember that r0 is a constant), and set this equal to zero (determine, then maximize the force function). Solve this equation for r in terms of r0 . The other constant should drop out. Person 2: Use Eq. (1.16) with the values of m and n of the Lennard-Jones potential to solve for the constant b in terms of a and the equilibrium bond distance, r0 . Now perform the determination of Fmax as given by Eq. (1.15); substitute this value of b back into Eq. (1.12), differentiate it twice with respect to r (remember that r0 is a constant), and set this equal to zero (determine, then maximize the force function). Solve this equation for r in terms of r0 . The other constant should drop out. Compare your answers. You should both get the same result. Answer : r = 1.1087r0 U

r

(a)

U

r

(b)

Figure 1.4 Schematic representation of the relationship between the shape of the potential energy well and selected physical properties. Materials with a deep well (a) have a high melting point, high elastic modulus, and low thermal expansion coefficient. Those with a shallow well (b) have a low melting point, low elastic modulus, and high thermal expansion coefficient. Adapted from C. R. Barrett, W. D. Nix, and A. S. Tetelman, The Principles of Engineering Materials. Copyright  1973 by Prentice-Hall, Inc.

17

INTRODUCTION AND OBJECTIVES

1.0.4.1 The Ionic Bond. To form an ionic bond, we must account for the complete transfer of electrons from one atom to the other. The easiest approach is to first transfer the electrons to form ions, then bring the two ions together to form a bond. Sodium chloride is a simple example that allows us to obtain both the bond energy and equilibrium bond distance using the potential energy approach. For this case, the potential energy, U , not only is the sum of the attractive and repulsive energies, UA and UR , respectively, but must also take into account the energy required to form ions from sodium and chlorine atoms, Eions . So, our energy expression looks like:

U = UA + UR + Eions

(1.17)

which at the equilibrium bond distance gives the equilibrium potential energy, U0 : U0 = UA,0 + UR,0 + Eions

(1.18)

Let us examine each of the three energies in Eq. (1.18) individually. Energy is required to form an ion of Na+ from elemental sodium. From Section 1.0.2.1, we already know that this process of removing an electron from an isolated atom is the ionization energy, IE, which for sodium is 498 kJ/mol. Similarly, energy is given off when Cl− is formed by adding an electron to an isolated gaseous atom of chlorine. This is the electron affinity, EA (see Section 1.0.2.2), which for chlorine is −354 kJ/mol. So, the energy required to form ions is given by: Eions = IE Na + EACl = 498 − 354 = 144 kJ/mol

(1.19)

For a diatomic molecule, the attraction between the two ions is strictly due to opposite charges, so the attractive force is given by Coulomb’s Law : FA = (Z1 e × Z2 e)/(4πε0 r 2 )

(1.20)

where εo is a constant called the electric permittivity (8.854 × 10−12 C2 /N · m2 ), e is the charge of an electron (1.6 × 10−19 C), Z is respective numbers of charge of positive and negative ions (Z1 = Z2 = 1), and r is the separation distance between ions in meters. (We will learn more of the electric permittivity in Chapter 6.) Substituting the values of Z in for sodium and chloride ions gives FA = +e2 /(4πε0 r 2 )

(1.21)

Energy is released by bringing the ions from infinite separation to their equilibrium separation distance, r0 . Recall that energy and force are related to one another by Eq. (1.13), so that the equilibrium attractive energy, UA,0 , can be found by integrating FA :  r0

UA,0 = −



FA dr = e2 /(4πε0 r0 )

(1.22)

Note the similarity in form of this expression for the attractive energy with that of Eq. (1.9). The exponent on r is 1, as it should be for ions (m = 1), and the other

18

THE STRUCTURE OF MATERIALS

parameters can be grouped to form the constant, a. Recall that by definition the attractive energy is a negative energy, so we will end up inserting a minus sign in front of the expression in Eq. (1.22) to match the form shown in Eq. (1.9). The repulsive energy can be derived simply by using the general expression given in Eq. (1.10) and solving for the constant b by minimizing the potential energy function [Eq. (1.11)] with a knowledge of the constant a from Eq. (1.22) and (1.9) (you should try this) to obtain: UR,0 = e2 /(4πε0 nr0 ) (1.23) where e and ε0 are the same as for the attractive force, r0 is again the equilibrium separation distance, and n is the repulsion exponent. Inserting UA,0 and UR,0 into the main energy expression, Eq. (1.18), (recall that the attractive energy must be negative) gives us the equilibrium potential energy, U0 : e2 −e2 + + Eions 4πε0 r0 4πε0 nr0

(1.24)

   1 −e2 + Eions U0 = 1 − n 4πε0 r0

(1.25)

U0 = and simplifying gives:

We solved for the equilibrium bond distance, r0 , in Eq. (1.16), and the constants a and b have, in effect, just been evaluated. Inserting these values into Eq. (1.25), along with Eq. (1.19) and using n = 8 (why?), gives: 

1 U0 = 1 − 8



 1(1.6 × 10−19 C)2 (6.02 × 1023 mol−1 ) + 142 4π(8.854 × 10−12 C2 /N · m2 )(2.36 × 10−10 m)

= −371 kJ/mol Similarly, we can calculate bond energies for any type of bond we wish to create. Refer to Appendix 1 for bond energy values. When we have an ordered assembly of atoms called a lattice, there is more than one bond per atom, and we must take into account interactions with adjacent atoms that result in an increased interionic spacing compared to an isolated atom. We do this with the Madelung constant, αM . This parameter depends on the structure of the ionic crystal, the charge on the ions, and the relative size of the ions. The Madelung constant fits directly into the energy expression (Eq. 1.25):    1 −e2 UL = αM − + Eions n 4πε0 r0

(1.26)

For NaCl, αM = 1.75 so the lattice energy, UL , is −811 kJ/mol. Typical values of the Madelung constant are given in Table 1.7 for different crystal structures (see Section 1.1.1). In general, the lattice energy increases (becomes more negative) with decreasing interionic distance for ions with the same charge. This increase in lattice energy translates directly into an increased melting point. For example, if we replace

INTRODUCTION AND OBJECTIVES

Table 1.7

19

Typical Values for the Madelung Constant

Compound

Crystal Lattice (see Section 1.1.1)

NaCl CsCl CaF2 CdCl2 MgF2 ZnS (wurtzite) TiO2 (rutile) β-SiO2

FCC CsCl Cubic Hexagonal Tetragonal Hexagonal Tetragonal Hexagonal

αM 1.74756 1.76267 2.51939 2.244 2.381 1.64132 2.408 2.2197

the chlorine in sodium chloride with other halogens, while retaining the cubic structure, the interionic spacing should change, as well as the melting point. We could also account for additional van der Waals interactions, but this effect is relatively small in lattices. 1.0.4.2 The Covalent Bond. Recall that covalent bonding results when electrons are “shared” by similar atoms. The simplest example is that of a hydrogen molecule, H2 . We begin by using molecular orbital theory to represent the bonding. Two atomic orbitals (1s) overlap to form two molecular orbitals (MOs), represented by σ : one bonding orbital (σ 1s), and one antibonding orbital, (σ ∗ 1s), where the asterisk superscript indicates antibonding. The antibonding orbitals are higher in energy than corresponding bonding orbitals. The shapes of the electron cloud densities for various MOs are shown in Figure 1.5. The overlap of two s orbitals results in one σ -bonding orbital and one σ -antibonding orbital. When two p orbitals overlap in an end-to-end fashion, as in Figure 1.5b, they are interacting in a manner similar to s –s overlap, so one σ -bonding orbital and one σ -antibonding orbital once again are the result. Note that all σ orbitals are symmetric about a plane between the two atoms. Side-to-side overlap of p orbitals results in one π-bonding orbital and one π-antibonding orbital. There are a total of four π orbitals: two for px and two for py . Note that there is one more node (region of zero electron density) in an antibonding orbital than in the corresponding bonding orbital. This is what makes them higher in energy. As in the case of ionic bonding, we use a potential energy diagram to show how orbitals form as atoms approach each other, as shown in Figure 1.6. The electrons from the isolated atoms are then placed in the MOs from bottom to top. As long as the number of bonding electrons is greater than the number of antibonding electrons, the molecule is stable. For atoms with p and d orbitals, diagrams become more complex but the principles are the same. In all cases, there are the same number of molecular orbitals as atomic orbitals. Be aware that there is some change in the relative energies of the π and σ orbitals as we go down the periodic chart, particularly around O2 . As a result, you might see diagrams that have the π2p orbitals lower in energy than the σ 2p. Do not let this confuse you if you see some variation in the order of these orbitals in other texts or references. For our purposes, it will not affect whether the molecule is stable or not.

20

THE STRUCTURE OF MATERIALS

Atomic orbitals

Molecular orbitals

(a )

+ A +

+

2s

2s

− A

( b)

s bonding

B

B

+

s* antibonding

+ A +



+

2p x

s bonding

B



2px

− A

(c )

B

+

s* antibonding

+ A +

+





2p y

2py

p bonding

B −

+

− A



B +

p* antibonding

Figure 1.5 The shape of selected molecular orbitals formed from the overlap of two atomic orbitals. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Antibonding molecular orbital

Energy

s*

Atomic orbital

Atomic orbital

1s

1s

Atom A

s

Atom B

Bonding molecular orbital Molecule

Figure 1.6 Molecular orbital diagram for the hydrogen molecule, H2 . Reprinted, by permission, from R. E. Dickerson, H. B. Gray, and G. P. Haight, Jr., Chemical Principles, 3rd ed., p. 446. Copyright  1979 by Pearson Education, Inc.

INTRODUCTION AND OBJECTIVES

21

s*2p p*2p 2p

2p

p2p

s2p s*2s 2s

2s s2s

O

O2

O

Figure 1.7 Molecular orbital diagram for molecular oxygen, O2 . From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Let us use molecular oxygen, O2 , as an example. As shown in Figure 1.7, each oxygen atom brings six outer-core electrons to the molecular orbitals. (Note that the 1s orbitals are not involved in bonding, and are thus not shown. They could be shown on the diagram, but would be at a very low relative energy at the bottom of the diagram.) The 12 total electrons in the molecule are placed in the MOs from bottom to top; according to Hund’s rule, the last two electrons must be placed in separate π ∗ 2p orbitals before they can be paired. The pairing of electrons in the MOs can manifest itself in certain physical properties of the molecule. Paramagnetism results when there are unpaired electrons in the molecular orbitals. Paramagnetic molecules magnetize in magnetic fields due to the alignment of unpaired electrons. Diamagnetism occurs when there are all paired electrons in the MOs. We will revisit these properties in Chapter 6. We can use molecular orbital theory to explain simple heteronuclear diatomic molecules, as well. A molecule such as hydrogen fluoride, HF, has molecular orbitals, but we must remember that the atomic orbitals of the isolated atoms have much different energies from each other to begin with. How do we know where these energies are relative to one another? Look back at the ionization energies in Table 1.4, and you see that the first ionization energy for hydrogen is 1310 kJ/mol, whereas for fluorine it is 1682 kJ/mol. This means that the outer-shell electrons have energies of −1310 (1s electron) and −1682 kJ/mol (2p electron), respectively. So, the electrons in fluorine are more stable (as we would expect for an atom with a much larger nucleus relative to hydrogen), and we can construct a relative molecular energy diagram for HF (see Figure 1.8) This is a case where the electronegativity of the atoms is useful. It qualitatively describes the relative energies of the atomic orbitals and the shape of the resulting MOs. The molecular energy level diagram for the general case of molecule AB where B is more electronegative than A is shown in Figure 1.9, and the corresponding molecular orbitals are shown in Figure 1.10. In Figure 1.9, note how the B atomic orbitals are lower in energy than those of atom A. In Figure 1.10, note how the number of nodes increases from bonding to antibonding orbitals, and also note how the electron probability is greatest around the more electronegative atom.

22

THE STRUCTURE OF MATERIALS

Example Problem 1.2

Construct the molecular orbital diagram for diatomic nitrogen, N2 . Is this molecule paramagnetic or diamagnetic? Answer: Diatomic nitrogen has all paired electrons, so it is diamagnetic.

N2 σ*2p π*2p 2p

2p

π2p

σ2p σ*2s 2s

2s σ2s

H orbital

HF orbitals s*

F orbitals

Energy

1s

px

py

2p

s

2s

2s

Figure 1.8 Molecular orbital diagram for HF. Reprinted, by permission, from R. E. Dickerson, H. B. Gray, and G. P. Haight, Jr., Chemical Principles, 3rd ed., p. 461. Copyright  1979 by Pearson Education, Inc.

INTRODUCTION AND OBJECTIVES

A orbitals

AB orbitals s*z p*x

23

B orbitals

p*y

np np Energy

sz px

py s*s

ns ns ss

Figure 1.9 Molecular orbital diagram for the general case of a diatomic molecule AB, where B is more electronegative than A. Reprinted, by permission, from R. E. Dickerson, H. B. Gray, and G. P. Haight, Jr., Chemical Principles, 3rd ed., p. 464. Copyright  1979 by Pearson Education, Inc.

Molecular orbitals don’t explain everything and become increasingly more difficult to draw with more than two atoms. We use a model called hybridization to explain other effects, particularly in carbon compounds. Hybridization is a “mixing” of atomic orbitals to create new orbitals that have a geometry better suited to a particular type of bonding environment. For example, in the formation of the compound BeH2 , we would like to be able to explain why this molecule is linear; that is, the H–Be–H bond is 180◦ .

A

B

z

(a) 0 + − A

B

z

(b)

Figure 1.10 The shape of selected molecular orbitals for the diatomic molecule AB, where B is more electronegative than A: (a) σ , (b) σ ∗ , (c) π and (d) π ∗ .

24

THE STRUCTURE OF MATERIALS

x

x

+

A

B

0

z



(c)

x

x 0

+



A

0

B



z

+

(d)

Figure 1.10 (continued ).

The hydrogen atoms only have one electron each to donate, both in their respective 1s orbitals, but beryllium has two electrons in the 2s orbital, and because its principal quantum number is two, it also has 2p orbitals, even though they are empty. The trick is to make two equivalent orbitals in Be out of the atomic orbitals so that each hydrogen will see essentially the same electronic environment. We can accomplish this by “mixing” the 2s orbital and one of the empty 2p orbitals (say, the 2pz ) to form two equivalent orbitals we call “sp” hybrids, since they have both s and p characteristics. As with molecular orbital theory, we have to end up with the same number of orbitals we started with. The bonding lobes on the new spa and spb orbitals on Be are 180◦ apart, just as we need to form BeH2 . In this manner, we can mix any type of orbitals we wish to come up with specific bond angles and numbers of equivalent orbitals. The most common combinations are sp, sp 2 , and sp 3 hybrids. In sp hybrids, one s and one p orbital are mixed to get two sp orbitals, both of which

INTRODUCTION AND OBJECTIVES

25

are 180◦ apart. A linear molecule results e.g., BeH2 , as shown in Figure 1.11. In sp 2 hybrids, one s and two p orbitals are mixed to obtain three sp 2 orbitals. Each orbital has 1/3 s and 2/3 p characteristic. A trigonal planar orbital arrangement results, with 120◦ bond angles. An example of a trigonal planar molecule is BF3 , as in Figure 1.12. Finally, in sp 3 hybrids, when one s and three p orbitals are mixed, four sp 3 orbitals result, each having 1/4 s and 3/4 p characteristic. The tetrahedral arrangement of orbitals creates a 109.5◦ bond angle, as is found in methane, CH4 (Figure 1.13). The concept of hybridization not only gives us a simple model for determining the correct geometry in simple molecules, but also provides us with a rationalization for multiple bonds. A double bond can result from sp 2 hybridization: one sp 2 –sp 2 bond and one π bond that forms between the p orbitals not involved in hybridization. An example is in C2 H4 (ethylene, Figure 1.14a), where each carbon undergoes sp 2 hybridization so that it can form an sp 2 –1s bond with two hydrogens and an sp 2 –sp 2 bond with the other carbon. The remaining p orbitals on each carbon (say, pz ) share electrons, which form the C–C double bond. A triple bond can be explained in terms of sp hybridization. It is formed from one sp–sp bond and two π bonds which form

Be

H

H

Figure 1.11 The linear structure of BeH2 .

F

F

B

F

The trigonal planar structure of BF3 .

Figure 1.12

H H C

H H

Figure 1.13 The tetrahedral structure of CH4 .

26

THE STRUCTURE OF MATERIALS

H 1s H 1s

• •

sp 2



sp 2

C

sp 2

• •

sp 2

C

sp 2 sp 2





1s H

• • •

1s H

(a)

H 1s

• •

sp

C

sp

• •

sp

C

sp

• •

1s

H

(b)

Figure 1.14 Hybridization resulting in (a) double bond and (b) triple bond.

between the two remaining p orbitals after hybridization. Acetylene (Figure 1.14b), C2 H2 , is such a compound in which both carbons undergo sp hybridization so that they can accommodate one bond with each other and one with hydrogen. Bonds can form between the remaining p orbitals, which in this case could be the pz and py orbitals on each carbon, for a total of three bonds between the carbon atoms.

Example Problem 1.3

Consider the molecule NF3 . How can we explain the observation that the F–N–F bond angles in NF3 are 107.3◦ and not 120◦ , as we might predict? Answer: Nitrogen undergoes sp 3 hybridization, not sp 2 , so it is tetrahedral. The additional 3 sp orbital is occupied by a lone pair of electrons from the nitrogen. This lone pair results in electron–electron repulsion that causes the other sp 3 orbitals bonded to fluorines to be closer together than the normal 109◦ tetrahedral bond angle, hence the 107.3 F–N–F bond angle.

1.0.4.3 The Metallic Bond. Some elements, namely those in the first two columns of the periodic table (IA and IIA) and the transition metals (e.g., Co, Mn, Ni), not only have a propensity for two atoms to share electrons such as in a covalent bond, but also have a tendency for large groups of atoms to come together and share valence

INTRODUCTION AND OBJECTIVES

27

electrons in what is called a metallic bond. In this case, there are now N atoms in the lattice, where N is a large number. There are N atomic orbitals, so there must be N MOs, many of which are degenerate, or of the same energy. This leads to bands of electrons, as illustrated in Figure 1.15 for sodium. The characteristics of the metallic bond are that the valence electrons are not associated with any particular atom in the lattice, so they form what is figuratively referred to as an electron gas around the solid core of metallic nuclei. As a result, the bonds in metals are nondirectional, unlike covalent or ionic bonds in which the electrons have a finite probability of being around a particular atom The electrons not involved in bonding remain in what is called the core band, whereas the valence electrons that form the electron gas enter into the valence band. 0

0 4s

3d

3p −10

3s

−10

Observed equilibrium value, r0 = 0.366 nm

1019 × E (J)

−30

−20

E (eV)

−20

−40

−30

−50 2p

−60

0

0.5

1.0

1.5

Interatomic spacing (nm)

Figure 1.15 Energy band diagram for a sodium lattice. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc. After J. C. Slater, Phys. Rev., 45, 794 (1934).

28

THE STRUCTURE OF MATERIALS

The remaining unfilled orbitals form higher-energy bands, called the conduction band. Keep in mind that even though the d and f orbitals may not be filled with electrons, they still exist for many of the heavier elements, so they must be included in the molecular orbital diagram. We will see later in Chapter 6 that the conduction band plays a very important role in the electrical, thermal, and optical properties of metals.

1.1

STRUCTURE OF METALS AND ALLOYS

Since the electrons in a metallic lattice are in a “gas,” we must use the core electrons and nuclei to determine the structure in metals. This will be true of most solids we will describe, regardless of the type of bonding, since the electrons occupy such a small volume compared to the nucleus. For ease of visualization, we consider the atomic cores to be hard spheres. Because the electrons are delocalized, there is little in the way of electronic hindrance to restrict the number of neighbors a metallic atom may have. As a result, the atoms tend to pack in a close-packed arrangement, or one in which the maximum number of nearest neighbors (atoms directly in contact) is satisfied. Refer to Figure 1.16. The most hard spheres one can place in the plane around a central sphere is six, regardless of the size of the spheres (remember that all of the spheres are the same size). You can then place three spheres in contact with the central sphere both above and below the plane containing the central sphere. This results in a total of 12 nearest-neighbor spheres in contact with the central sphere in the close-packed structure. Closer inspection of Figure 1.16a shows that there are two different ways to place the three nearest neighbors above the original plane of hard spheres. They can be directly aligned with the layer below in an ABA type of structure, or they can be rotated so that the top layer does not align core centers with the bottom layer, resulting in an ABC structure. This leads to two different types of close-packed structures. The ABAB. . . structure (Figure 1.16b) is called hexagonal close-packed (HCP) and the ABCABC. . . structure is called face-centered cubic (FCC). Remember that both

A

(a)

(b)

A

C

B

B

A

A

(c)

Figure 1.16 Close-packing of spheres. (a) Top view, (b) side view of ABA structure, (c) side view of ABC structure. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

STRUCTURE OF METALS AND ALLOYS

Figure 1.17

29

The extended unit cell of the hexagonal close-packed (HCP) structure.

of these close-packed arrangements have a coordination number (number of nearest neighbors surrounding an atom) of 12: 6 in plane, 3 above, and 3 below.∗ Keep in mind that for close-packed structures, the atoms touch each other in all directions, and all nearest neighbors are equivalent. Let us first examine the HCP structure. Figure 1.17 is a section of the HCP lattice, from which you should be able to see both hexagons formed at the top and bottom of what is called the unit cell. You should also be able to identify the ABA layered structure in the HCP unit cell of Figure 1.17 through comparison with Figure 1.16. Let us count the number of atoms in the HCP unit cell. The three atoms in the center of the cell are completely enclosed. The atoms on the faces, however, are shared with adjacent cells in the lattice, which extends to infinity. The center atoms on each face are shared with one other HCP unit cell, either above (for the top face) or below (for the bottom face), so they contribute only half of an atom each to the HCP unit cell under consideration. This leaves the six corner atoms on each face (12 total) unaccounted for. These corner atoms are at the intersection of a total of six HCP unit cells (you should convince yourself of this!), so each corner atom contributes only one-sixth of an atom to our isolated HCP unit cell. So, the total number of whole atoms in the HCP unit cell is 3×1

= 3 center atoms

2 × (1/2) = 1 face atom 12 × (1/6) = 2 corner atoms 6 total atoms ∗

At this point, you may find it useful to get some styrofoam spheres or hard balls to help you visualize some of the structures we will describe. We will use a number of perspectives, views, and diagrams to build these structures. Some will treat the atoms as solids spheres that can touch each other, some will use dots at the center of the atoms to help you visualize the larger structure. Not all descriptions will help you—find those perspectives that work best for you.

30

THE STRUCTURE OF MATERIALS

(a)

(b)

Figure 1.18 The face-centered cubic (FCC) structure showing (a) atoms touching and (b) atoms as small spheres. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 32. Copyright  2000 by John Wiley & Sons, Inc.

Counting the atoms in the FCC structure is performed in a similar manner, except that visualizing the FCC structure takes a little bit of imagination and is virtually impossible to show on a two-dimensional page. Take the ABC close-packed structure shown in Figure 1.16c, and pick three atoms along a diagonal. These three atoms form the diagonal on the face of the FCC unit cell, which is shown in Figure 1.18. There is a trade-off in doing this: It is now difficult to see the close-packed layers in the FCC structure, but it is much easier to see the cubic structure (note that all the edges of the faces have the same length), and it is easier to count the total number of atoms in the FCC cell. In a manner similar to counting atoms in the HCP cell, we see that there are zero atoms completely enclosed by the FCC unit cell, six face atoms that are each shared with an adjacent unit cell, and eight corner atoms at the intersection of eight unit cells to give 6 × (1/2) = 3 face atoms 8 × (1/8) = 1 corner atom 4 total atoms Remember that both HCP and FCC are close-packed structures and that each has a coordination number of 12, but that their respective unit cells contain 6 and 4 total atoms. We will now see how these two special close-packed structures fit into a larger assembly of crystal systems. 1.1.1

Crystal Structures

Our description of atomic packing leads naturally into crystal structures. While some of the simpler structures are used by metals, these structures can be employed by heteronuclear structures, as well. We have already discussed FCC and HCP, but there are 12 other types of crystal structures, for a total of 14 space lattices or Bravais lattices. These 14 space lattices belong to more general classifications called crystal systems, of which there are seven.

STRUCTURE OF METALS AND ALLOYS

31

z

c

α

β

b

y

γ a

x

Figure 1.19 Definition of a coordinate system for crystal structures.

Before describing each of the space lattices, we need to define a coordinate system. The easiest coordinate system to use depends upon which crystal system we are looking at. In other words, the coordinate axes are not necessarily orthogonal and are defined by the unit cell. This may seem a bit confusing, but it simplifies the description of cell parameters for those systems that do not have crystal faces at right angles to one another. Refer to Figure 1.19. For each crystal system, we will define the space lattice in terms of three axes, x, y, and z, with interaxial angles α, β, γ . Note that the interaxial angle α is defined by the angle formed between axes z and y, and also note that angles β and γ are defined similarly. Only in special circumstances are α, β, γ equal to 90◦ . The distance along the y axis from the origin to the edge of the unit cell is called the lattice translation vector, b. Lattice translation vectors a and c are defined similarly along the axes x and z, respectively. The magnitudes (lengths) of the lattice translation vectors are called the lattice parameters, a, b, and c. We will now examine each of the seven crystal systems in detail. 1.1.1.1 Crystal Systems. The cubic crystal system is composed of three space lattices, or unit cells, one of which we have already studied: simple cubic (SC), bodycentered cubic (BCC), and face-centered cubic (FCC). The conditions for a crystal to be considered part of the cubic system are that the lattice parameters be the same (so there is really only one lattice parameter, a) and that the interaxial angles all be 90◦ . The simple cubic structure, sometimes called the rock salt structure because it is the structure of rock salt (NaCl), is not a close-packed structure (see Figure 1.20). In fact, it contains about 48% void space; and as a result, it is not a very dense structure. The large space in the center of the SC structure is called an interstitial site, which is a vacant position between atoms that can be occupied by a small impurity atom or alloying element. In this case, the interstitial site is surrounded by eight atoms. All eight atoms in SC are equivalent and are located at the intersection of eight adjacent unit cells, so that there are 8 × (1/8) = 1 total atoms in the SC unit cell. Notice that

32

THE STRUCTURE OF MATERIALS

Crystal Structure

Cubic

Lattice Interaxial Parameters Angles

a=b=c

a=b=g= 90°

Simple

Orthorhombic

a≠b≠c

Rhombohedral

a=b=c

a = b = g ≠ 90°, < 120 °

Tetragonal

a=b≠c

a = b = g = 90°

a≠b≠c

Base-centered

Body-centered

Simple

Body-centered

a = g = 90°, b ≠ 90°

Simple

Triclinic

Hexagonal

a≠b≠c

a = b, a ≠ c

Face-centered

a=b=g= 90°

Simple

Monoclinic

Body-centered

Base-centered

a = b = g ≠ 90°

a = b = 90°, g = 120°

Figure 1.20 Summary of the 14 Bravais space lattices.

Face-centered

STRUCTURE OF METALS AND ALLOYS

33

Figure 1.20 shows the atoms as points, but the atoms actually occupy a larger space than that. In fact, for SC, the atoms touch along the edge of the crystal. Body-centered cubic (BCC) is the unit cell of many metals and, like SC, is not a close-packed structure. The number of atoms in the BCC unit cell are calculated as follows: 1×1

= 1 center atom

8 × (1/8) = 1 corner atom 2 total atoms Finally, face-centered cubic (FCC) has already been described (Figure 1.18). Even though FCC is a close-packed structure, there are interstitial sites, just as in SC. There are actually two different types of interstitial sites in FCC, depending on how many atoms surround the interstitial site. A group of four atoms forms a tetrahedral interstice, as shown in Figure 1.21. A group of six atoms arranged in an octahedron (an eight-sided geometric figure), creates an octahedral interstice (Figure 1.22). Figure 1.23 shows the locations of these interstitial sites within the FCC lattice. Note that there are eight total tetrahedral interstitial sites in FCC and there are four total octahedral interstitial sites in FCC (prove it!), which are counted in much the same way as we previously counted the total number of atoms in a unit cell. We will see later on that these interstitial sites play an important role in determining solubility of impurities and phase stability of alloys. Interstitial sites are the result of packing of the spheres. Recall from Figure 1.18 that the spheres touch along the face diagonal in FCC. Similarly, the spheres touch along the body diagonal in BCC and along an edge in SC. We should, then, be able to calculate the lattice parameter, a, or the length of a face edge, from a knowledge of the sphere radius. In SC, it should be evident that the side of a unit cell √is simply 2r. Application of a little geometry should prove to you that in FCC, a = 4r/ 2. The relationship between a and r for BCC is derived in Example Problem 1.4; other geometric relationships, including cell volume for cubic structures, are listed in Table 1.8. Finally, atomic radii for the elements can be found in Table 1.9. The radius of an atom is not an exactly defined quantity, and it can vary depending upon the bonding environment in which

Figure 1.21 A tetrahedral interstice. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

34

THE STRUCTURE OF MATERIALS

Figure 1.22 An octahedral interstice. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Figure 1.23 Location of interstitial sites in FCC. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

STRUCTURE OF METALS AND ALLOYS

35

Example Problem 1.4

˚ Calculate the lattice Molybdenum has a BCC structure with an atomic radius of 1.36 A. parameter for BCC Mo. Answer: We know H that the molybdenum b G atoms touch along the E E G r body diagonal in BCC, as shown in the projeca tion at right. The length 2r a of the body diagonal, then, is 4r, and is C related to the lattice r A a parameter, a (which is A C a a √2 the length of the cube B edge, not the length Reprinted, by permission, from Z. Jastrzebski, The Nature and of the face √ diagonal, Properties of Engineering Materials , p. 47, 2nd ed. Copyright  1976, John Wiley & Sons, Inc. which is a 2) by application of the Pythagorean theorem: √ (4r)2 = a 2 + (a 2)2 = 3a 2 √ √ a = 4r/ 3 = 4(1.36)/ 3 = 3.14 ˚ which is consistent with the value in Table 1.11. The lattice parameter for BCC Mo is 3.14 A,

Table 1.8 Summary of Important Parameters in the Cubic Space Lattices Simple Cubic Unit cell side, a Face diagonal Body diagonal Number of atoms Cell volume

√ 2r √2(2r) 3(2r) 1 8r 3

Face-Centered Cubic

Body-Centered Cubic

√ 4r/ 2 4r √ 3/2(4r) 4 32r 3 √ 2

√ 4r/ 3 √ 2/3(4r) 4r 2 64r 3 √ 3 3

r = atomic radius.

it finds itself. As a result, three types of radii are listed for each element in Table 1.9: an atomic radius of an isolated atom, an ionic radius, and a metallic radius. Just as in Figure 1.2 for electronic structure, there are some important trends in the atomic radii. The atomic radius tends to increase as one goes down the column in a series. This is due to the addition of energy levels and more electron density. Radii tend to decrease as we move across a row, because there is less shielding from inner electrons

2+ ions 4 Be 1.14 0.54 0.89 12 Mg 1.60 0.78 1.36 20 Ca 1.97 1.06 1.74 38 Sr 2.15 1.27 1.91 56 Ba 2.17 1.3 1.98

0.46 1.54 —

3 Li 1.52 0.78 1.23 11 Na 1.86 0.98 1.57 19 K 2.31 1.33 2.03 37 Rb 2.51 1.49 2.16 55 Cs 2.65 1.65 2.35

1 H

4+ ions 22 Ti 1.47 0.64 1.32 40 Zr 1.58 0.89 1.45 72 Hf 1.59 0.84 1.44

3+ ions 21 Sc 1.60 0.82+ 1.44 39 Y 1.81 1.06 1.62 57 La 1.87 1.22 1.69

Pb4− Pb2+ S6+ Te4+ Si4− Ti3+

23 V 1.32 0.45+ 1.22 41 Nb 1.43 0.74 1.34 73 Ta 1.47 0.68 1.34

5+ ions

2.15 1.32 0.34 0.89 1.98 0.69

24 Cr 1.25 0.3 1.18 42 Mo 1.36 0.65 1.30 74 W 1.37 0.68 1.30

6+ ions

Mn2+ Pt4+ Rh4+ V3+ Nb4+ Ti2+

25 Mn 1.12 0.52 1.17 43 Tc — — 1.27 75 Re 1.38 0.72 1.28

4+ ions

0.91 0.55 0.65 0.65 0.74 0.76

26 Fe 1.24 0.673+ 1.17 44 Ru 1.34 0.65 1.25 76 Os 1.35 0.67 1.26

4+ ions

Mn3+ Co2+ Mo4+ V4+ Tl+ Se6+

27 Co 1.25 0.65 1.16 45 Rh 1.34 0.68 1.25 77 Ir 1.35 0.664+ 1.27

3+ ions 28 Ni 1.25 0.78 1.15 46 Pd 1.37 0.50 1.28 78 Pt 1.38 0.52 1.30

2+ ions

0.70 Fe2+ 0.82 W6+ 0.68 Cr3+ 0.61 Sn4− 1.49 0.30–0.04

29 Cu 1.28 0.96 1.17 47 Ag 1.44 1.13 1.34 79 Au 1.44 1.37 1.34

1+ ions

30 Zn 1.33 0.83 1.25 48 Cd 1.50 1.03 1.41 80 Hg 1.50 1.12 1.44

2+ ions

0.87 0.65 0.64 2.15

Many elements have multiple valence states. Additional ionic radii are listed below.

Atomic, Ionic, and Metallic Radii of the Elements

5 B 0.97 0.2 0.81 13 Al 1.43 0.57 1.25 31 Ga 1.35 0.62 1.25 49 In 1.57 0.91 1.50 81 TI 1.71 1.49 1.55

3+ ions

˚ (1A ˚ = 0.1 nm). Source: Materials Science and Engineering Handbook; Pauling, Nature of the Chemical Bond. All values in angstroms, A

Atomic Ionic Metallic

Atomic Ionic Metallic

Atomic Ionic Metallic

Atomic Ionic Metallic

Atomic Ionic Metallic

Atomic Ionic Metallic

Table 1.9

6 C 0.77 447◦ C) FCC (>427◦ C) BCC (>1742◦ C) FCC (>912◦ C) BCC (>1394◦ C) BCC (< −193◦ C) BCC (< −233◦ C) Cubic (234◦ C) BCC (>883◦ C) BCC (>1481◦ C) BCC (>872◦ C)

For example, the center atom in the BCC space lattice (see Figure 1.20) has cell coordinates of 1/2, 1/2, 1/2. Any two points are equivalent if the fractional portions of their coordinates are equal: 1/2, 1/2, 1/2 ≡ −1/2, −1/2, −1/2 (center) 0, 0, 0 ≡ 1, 0, 1 (corner) A cell direction is designated by the vector r, which is a combination of the lattice translation vectors a, b, and c: r = ua + vb + wc

(1.27)

A direction can also be specified with the cell coordinates in square brackets, with commas and fractions removed: [1 1 1] ≡ [1/2 1/2 1/2] ≡ 1/2a + 1/2b + 1/2c Negative directions are indicated by an overbar [1-1] and are called the “one negative one one” direction. All directions are relative to the origin where the three lattice translation vectors originate (see Figure 1.19). The cell volume, V , can be calculated using the lattice translation vectors: V = |a × b · c|

(1.28)

Mathematically, this is a triple scalar product and can be used to calculate the volume of any cell, with only a knowledge of the lattice translation vectors. If the lattice parameters and interaxial angles are known, the following expression for V can be derived from the vector expression: V = abc(1 − cos2 α − cos2 β − cos2 γ + 2 cos α cos β cos γ )1/2

(1.29)

41

STRUCTURE OF METALS AND ALLOYS

This looks complicated, but for orthogonal systems (α, β, γ = 90◦ ) it reduces quite nicely to the expected expression: V = abc

(1.29a)

Now that we know how to find the cell volume, we can use some previous information to calculate an important property of a material, namely, its density, which we represent with the lowercase Greek letter rho, ρ. For example, aluminum has an FCC space lattice. Recall that there are four atoms in the FCC unit cell. We know that each aluminum atom has an atomic weight of 27 g/mol. From Table 1.11, the cubic lattice ˚ or 0.405 nm (4.05 × 10−8 cm). This gives us a parameter for aluminum is 4.05 A, 3 −23 3 cm . You should confirm that the theoretical density for volume of a = 6.64 × 10 aluminum is then: ρ=

(4 atoms/unit cell)(27 g/mol) = 2.70 g/cm3 per unit cell (6.02 × 1023 atoms/mol)(6.64 × 10−23 cm3 )

Cooperative Learning Exercise 1.2 The actinide-series element protactinium (Pa, AW = 231.04) has a body-centered tetragonal structure with cell dimensions a = 0.3925 nm, c = 0.3238 nm. Person 1: Determine the weight, in grams, of a single unit cell of Pa. Person 2: Calculate the volume of single Pa unit cell. Combine your answers appropriately to arrive at the density.

Answer : 15.38 g/cm3

In addition to cell coordinates and directions, crystal planes are very important for the determination and analysis of structure. We begin with the cell’s coordinate system, with axes x, y, and z. Recall that the axes are not necessarily orthogonal and that a, b, and c are the lattice parameters. Look at Figure 1.24. The equation of an arbitrary plane with intercepts A, B, and C, relative to the lattice parameters is given by 1x 1y 1z + + =1 Aa Bb Cc

(1.30)

We designate a plane by Miller indices, h, k, and l, which are simply the reciprocals of the intercepts, A, B, and C:   1 1 1 = (hkl) (1.31) ABC Note that the Miller indices are enclosed in parentheses and not separated by commas. Miller indices are determined as follows: ž ž ž

Remove all indeterminacy; that is, the planes should have nonzero intercepts. Find intercepts along three axes of the crystal system. Take the reciprocals of the intercepts in terms of a, b, and c.

42

THE STRUCTURE OF MATERIALS

z

y

C

B

g

O

D b a

A

x

Figure 1.24 Definition of Miller indices for an arbitrary plane (shaded area). From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

ž

ž

Multiply reciprocals by the smallest integer necessary to convert them into a set of integers. Enclose resulting integers in parentheses, (hkl), without commas.

Any planes that have common factors are parallel. For example, a (222) and a (111) plane are parallel, as are (442) and (221) planes. As with cell directions, a minus sign (in this case, indicating a negative intercept) is designated by an overbar. The (221) plane has intercepts at 1/2, −1/2, and 1 along the x, y, and z axes, respectively. Some important planes in the cubic crystal system are shown in Figure 1.25. In a manner similar to that used to calculate the density of a unit cell, we can calculate the density of atoms on a plane, or planar density. The perpendicular intersection of a plane and sphere is a circle, so the radius of the atoms will be helpful in calculating the area they occupy on the plane. Refer back to Example Problem 1.4 when we calculated the lattice parameter for a BCC metal. The section shown along the body diagonal is actually the (110) plane. The body-centered atom is entirely enclosed by this plane, and the corner atoms are located at the confluence of four adjacent planes, so each contributes 1/4 of an atom to the (110) plane. So, there are a total of two atoms on the (110) plane. If we know the lattice parameter or atomic radius, we can calculate the area of the plane, Ap , the area occupied by the atoms, Ac , and the corresponding

STRUCTURE OF METALS AND ALLOYS

(100)

(110)

43

(111)

Figure 1.25 Some important planes in the cubic space lattice.

planar density, PD = Ac /Ap . We will see in Chapter 5 that an important property of metals, ductility, is affected by planar density. 1.1.1.3 Interplanar Spacings. To this point, we have concentrated on planes in an isolated cell. A crystal lattice, of course, is composed of many individual unit cells, with the planes extending in all directions. So, a real crystal lattice has many (111) planes, for example, all of which are parallel to one another. There is a uniform distance between like planes in a lattice, which we call the interplanar spacing and designate with d, the perpendicular distance between adjacent planes in a set. Note that even though the (111) and (222) planes are parallel to one another, they are not the same plane, since their planar densities may be much different depending on the lattice (for example, compare these two planes in simple cubic). What we are calculating here is the perpendicular distance between the same plane in adjacent cells. The general expression for d in terms of lattice parameters and interaxial angles is somewhat complicated

−1/2 h2 b2 c2 sin2 α + k 2 a 2 c2 sin2 β + l 2 a 2 b2 sin2 γ d = V  + 2hlab2 c(cos α cos γ − cos β) + 2hkabc2 (cos α cos β − cos γ )  + 2kla 2 bc(cos β cos γ − cos α) (1.32) where V is the cell volume as defined by Eq. (1.28). Note that the Miller indices of the plane under consideration are included here, to distinguish the distance between different planes in the same space lattice. For orthogonal systems (orthorhombic, tetragonal, and cubic), α = β = γ = 90◦ 

1 = d



k2 l2 h2 + + a2 b2 c2

1/2

(1.33)

For a cubic system (a = b = c), this expression simplifies even further to d=√

a h2 + k 2 + l 2

(1.34)

We will see that the interplanar spacing is an important parameter for characterizing many types of materials.

44

THE STRUCTURE OF MATERIALS

Cooperative Learning Exercise 1.3

The (111) plane in FCC is shown at right. Person 1: Calculate the area of the (111) plane, Ap , in the FCC cell in terms of the atomic radius, r. Person 2: Calculate the area occupied by the atoms, Ac , in the (111) plane of the FCC cell in terms of r. Combine your answers appropriately to arrive at ˚ what is the the planar density, PD. If r = 1.36 A, number density (number of atoms per unit area) for the (111) plane in FCC?

(111)

(110)

a

Reprinted, by permission, from Z. Jastrzebski, The Nature and Properties of Engineering Materials, p. 37, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc.

Answer : P D = 0.91; number density = 1.72 × 1015 atoms/cm2

1.1.2

X-Ray Diffraction

The interplanar spacing between adjacent planes allows us to use a very powerful tool for structural determination called X-ray diffraction (XRD). If we bombard a crystal with X rays of a certain wavelength, λ, at an incident angle, θ , the X rays can either pass directly through the crystal or, depending on the angle of incidence, interact with certain atoms in the lattice. This interaction is best visualized as a “shooting gallery” in which the incident X rays “bounce off” the hard-core spheres in the lattice (see Figure 1.26). In reality, the X-ray photons are interacting with the electron density around the atoms, leading to diffraction. The path difference for rays reflected from adjacent planes is diffracted beam

incident beam

θ

θ

d

atomic planes

Figure 1.26 Schematic illustration of incident radiation diffraction by a crystal lattice.

STRUCTURE OF METALS AND ALLOYS

45

Intensity of diffracted beam

200

110 310

211

222

220 20

40

60

400

321

80

100

120

140

160

2q

Figure 1.27 X-Ray diffraction pattern for tungsten. Adapted from A. G. Guy, and J. J. Hren, Elements of Physical Metallurgy, p. 208, 3rd ed. Copyright  1974 by Addison-Wesley.

2d(sin θ ), and constructive interference occurs when the path difference is an integer, n, of the wavelength: nλ = 2d sin θ (1.35) Equation (1.35) is called Bragg’s law and is a very important result. It says that if we bombard a crystal lattice with X rays of a known wavelength and at a known angle, we will be able to detect diffracted X rays of various intensities that represent a specific interplanar spacing in the lattice. The result of scanning through different angles at a fixed wavelength and counting the number (intensity) of diffracted X rays is a diffraction pattern, an example of which is shown in Figure 1.27. Note how the sum of the Miller indices, or order, for each plane increases from left to right in the pattern, and also note how different planes have different intensities. Also, not all planes result in diffracted X rays. From the abscissa (which is by convention listed as 2θ for instrumentation reasons), the interplanar spacing, d, can be calculated using Bragg’s law. The study of crystals and their X-ray patterns is both fascinating and complex. From a known crystal structure, a theoretical diffraction pattern can be constructed using vector algebra, including such considerations as structure factors which account for disallowed interactions with the crystal lattice and determine the relative intensity of the diffracted beams. For our purposes, it is sufficient to know that a diffraction pattern can be obtained experimentally in a relatively routine fashion, and the resulting pattern is characteristic of a specific crystalline material. 1.1.3

Point Defects

Now that the most important aspects of perfect crystals have been described, it is time to recognize that things are not always perfect, even in the world of space lattices. This is not necessarily a bad thing. As we will see, many important materials phenomena that are based on defective structures can be exploited for very important uses. These defects, also known as imperfections, are grouped according to spatial extent.

46

THE STRUCTURE OF MATERIALS

Point defects have zero dimension; line defects, also known as dislocations, are onedimensional; and planar defects such as surface defects and grain boundary defects have two dimensions. These defects may occur individually or in combination. Let us first examine what happens to a crystal when we remove, add, or displace an atom in the lattice. We will then describe how a different atom, called an impurity (regardless of whether or not it is beneficial), can fit into an established lattice. As shown by Eq. (1.36), point defects have equilibrium concentrations that are determined by temperature, pressure, and composition. This is not true of all types of dimensional defects that we will study.   Nd = N exp −Ed kB T

(1.36)

In Eq. (1.36), Nd is the equilibrium number of point defects, N is the total number of atomic sites per volume or mole, Ed is the activation energy for formation of the defect, kB is Boltzmann’s constant (1.38 × 10−23 J/atom · K), and T is absolute temperature. Equation (1.36) is an Arrhenius-type expression of which we will see a great deal in subsequent chapters. Many of these Arrhenius expressions can be derived from the Gibbs free energy, G. When an atom is missing from a lattice, the resulting space is called a vacancy (not to be confused with a “hole,” which has an electronic connotation), as in Figure 1.28. In this case, the activation energy, Ed , is the energy required to remove an atom from the lattice and place it on the surface. The activation energy for the formation of vacancies in some representative elements is given in Table 1.13, as well as the corresponding vacancy concentration at various temperatures. Note that the vacancy concentration decreases at lower temperatures. In a nonequilibrium situation, such as rapid cooling from the melt, we would not expect the equilibrium concentration to be attained. This

Self-interstitial Vacancy

Figure 1.28 Representation of a vacancy and self-interstitial in a crystalline solid. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

47

STRUCTURE OF METALS AND ALLOYS

Table 1.13 Formation Energy of Vacancies for Selected Elements and Equilibrium Concentrations at Various Temperatures

Element Ag Al Au Cu Ge K Li Mg Na Pt Si

Ed (kJ/mol) 106.1 73.3 94.5 96.4 192.9 38.6 39.5 85.8 38.6 125.4 221.8

Melting Point, Tm (◦ C) 960 660 1063 1083 958 63 186 650 98 1769 1412

Nd (vacancies/cm3 ) 25◦ C

300◦ C 4

1.5×10 1.0×1010 1.5×106 1.1×106 19◦ C) α-Monoclinic β-Hexagonal Orthorhombic

7.417 5.66 6.65 19.08 14.50

4.945 — 20.96 — 5.60

2.547 19.50 6.50 6.49 7.40

1.00 2.30 0.936 0.922 0.93

Trigonal Orthorhombic Monoclinic Orthorhombic α-Monoclinic β-Orthorhombic Orthorhombic

21.90 10.6 7.81 8.57 4.96 8.58 20.98

— 5.4 2.25 4.95 9.64 4.91 12.06

6.65 5.1 5.51 2.52 4.62 2.56 10.40

1.13 1.42 1.35 1.430 1.925 1.973 1.26

Monoclinic Monoclinic Monoclinic triclinic Orthorhombic

8.63 4.60 8.05 4.71 10.46

9.11 9.50 13.04 4.44 4.66

4.83 8.60 19.48 7.12 7.03

1.04 1.01 1.228 1.197 1.126

α-Triclinic β-Triclinic Triclinic Triclinic Monoclinic Orthorhombic Orthorhombic Orthorhombic Orthorhombic Orthorhombic Trigonal

4.9 4.9 4.93 4.59 4.70 7.97 8.50 6.88 10.76 8.50 16.25

5.4 8.0 4.58 5.14 8.66 4.76 4.95 11.91 5.76 4.95 —

17.2 17.2 16.8 13.9 33.9 7.57 6.70 18.60 7.00 6.70 6.50

1.24 1.248 1.27 1.33 1.34 1.296 1.416 0.972 1.10 1.60 1.168

Source: Tadokoro, Structure of Crystalline Polymers.

at which the slope of the specific volume with temperature curve decreases (between C and D), representing solidification, is called the glass transition temperature, Tg . The reasoning behind the term “glass transition temperature” becomes more apparent if we turn around and begin slowly heating the glassy polymer. At some point, there is sufficient mobility in the polymer chains for them to begin to align themselves in a regular array and form crystallites. The polymer is not yet molten—there is simply short range chain movement that results in an amorphous to crystalline transformation in the solid state. This point is also Tg . As we continue to heat the sample, the nowcrystalline polymer eventually reaches Tm and melts. Most polymers are a combination

STRUCTURE OF POLYMERS

93

Intensity

110

Amorphous

200

Background Diffraction angle

Figure 1.66 Resolution of the X-ray scattering curve of a semicrystalline polyethylene sample into contributions from crystalline (110 and 200 planes) and amorphous components. From F. W. Billmeyer, Textbook of Polymer Science, 3rd ed. Copyright  1984 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

of amorphous and crystalline components and are the result of intermediate cooling paths such as ABEF in Figure 1.67. Keep in mind that some polymers, no matter how slowly we cool them, cannot crystallize, and they follow the glassy path ABCD all the time. We will have more to say about the glass transition temperature in Section 1.3.7. 1.3.6.2 Liquid Crystalline Polymers. One class of polymers that requires some special attention from a structural standpoint is liquid crystalline polymers, or LCPs. Liquid crystalline polymers are nonisotropic materials that are composed of long molecules parallel to each other in large clusters and that have properties intermediate between those of crystalline solids and liquids. Because they are neither completely liquids nor solids, LCPs are called mesophase (intermediate phase) materials. These mesophase materials have liquid-like properties, so that they can flow; but under certain conditions, they also have long-range order and crystal structures. Because they are liquid-like, LCPs have a translational degree of freedom that most solid crystals we have described so far do not have. That is, crystals have three-dimensional order, whereas LCPs have only one- or two-dimensional order. Nevertheless, they are called “crystals,” and we shall treat them as such in this section. In many cases, these polymer chains take on a rod-like (calamitic LCPs) or even disc-like (discotic LCPs) conformation, but this does not affect the overall structural classification scheme. There are many organic compounds, though not polymeric in nature, that exhibit liquid crystallinity and play important roles in biological processes. For example, arteriosclerosis is possibly caused by the formation of a cholesterol containing liquid crystal in the arteries of the heart. Similarly, cell wall membranes are generally considered to have liquid crystalline properties. As interesting as these examples of liquid crystallinity in small, organic compounds are, we must limit the current discussion to polymers only.

94

THE STRUCTURE OF MATERIALS

A B C D E V

F

G

Tg

Tm T

Figure 1.67 Specific volume as a function of temperature on cooling from the melt for a polymer that tends to crystallize. Region A is liquid, B liquid with elastic response, C supercooled liquid, D glass, E crystallites in a supercooled liquid matrix, F crystallites in a glassy matrix, and G completely crystalline. Paths ABCD, ABEF, and ABG represent fast, intermediate, and very slow cooling rates, respectively. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

There are three categories of LCPs, grouped according to the arrangement of the molecules: smectic, nematic, and cholesteric. Nematic (from the Greek term meaning “thread-like”) LCPs have their molecules aligned along the chain axis, as shown schematically in Figure 1.68. Nematic liquids have low viscosity, and tend to be turbid, or “cloudy.” Smectic (from the Greek term for “soap”) LCPs have an additional level of structure, in that the polymer chains are also aligned along the chain axis, but they also segregate into layers. Smectic liquids are also turbid, but tend to be highly viscous. Finally, cholesteric (from the Greek term for “bile”) LCPs have layered structures, but the aligned polymer chains in one layer are rotated from the aligned polymer chains in adjacent layers. Cholesteric LCPs are also highly viscous, but often possess novel photochromic, optical, thermochromic, and electro-optical properties. Clearly, not all polymeric molecules possess the ability to form LCPs. Generally, LCPs have a molecular structure in which there are two regions with dissimilar chemical properties. For example, chains that consist of aliphatic–aromatic, dipolar–nonpolar, hydrophobic–hydrophilic, flexible–rigid, or hydrocarbon–fluorocarbon combinations of substantial size have a propensity to form LCPs. The two different portions of the chains can interact locally with similar regions of adjacent chains, leading to ordering. Liquid crystalline polymers that rely on structural units in the backbone, or main chain, to impart their crystallinity are called main-chain LCPs. The building blocks of typical main-chain LCPs are shown in Figure 1.69. Side-chain LCPs crystallize due to interaction of the side chains, or branches, as shown in Figure 1.69. Side

STRUCTURE OF POLYMERS

95

(a)

(b)

Figure 1.68 The structure of liquid crystalline polymers (a) nematic, (b) smectic and (c) cholesteric. Reprinted, by permission, from J. L. Fergason, Scientific American, 211(2), pp. 78, 80. Copyright  1964 by Scientific American, Inc.

96

THE STRUCTURE OF MATERIALS

(c)

Figure 1.68 (continued ).

chains play an important role in determining not only the liquid crystalline activity of polymers, but the type of structure that will form as well. For example, in Table 1.25, we see that the number of hydrocarbon units in the side chain (not the backbone repeat unit) affects whether the resulting LCP is nematic or smectic. We have used the general term “liquid” to describe this special class of polymers, but we know that a liquid can be either a melt or a solution. In the case of LCPs, both

STRUCTURE OF POLYMERS

97

Table 1.25 Effect of the Flexible Tail on the Structure of a Side-Chain LCP CH3

C

COO

(CH2)n

O

COO

R

CH2

S.C. LCPs

Number

n

R

Structure

1 2 3 4

2 2 6 6

OCH3 OC3 H7 OCH3 OC6 H13

Nematic Smectic Nematic Smectic

Flexible Tail

M.C. LCPs Flexible Tail

Cyclic Unit

Bridging Group

Cyclic Unit

Functional Unit

Spacer

Cyclic Unit

Bridging Group

Cyclic Unit

Functional Unit

Spacer

Cyclic Unit

none R

1,3 1,4

X=Me Ph Cl

X

Functional Unit

none

none

CO

OR CN

Bridging Group

O CR

CO

CR

NO

O

CO

NH

NO

N

C

C

CR

N

n Flexible Backbone CH

none (CH2)n O CO

S

R

SiR2

S

NR'

R

SiR

CHR O

O

(CH2 CHR)n

N

Flexible Backbone n

Spacer

O

CR

Functional Unit

SiR

O

SiR2

O

NR'

CR

1,4 1,5 2,6

n=1,2,3

Cholesteryl

Figure 1.69 General structure of main-chain (M.C.) and side-chain (S.C.) LCPs. Adapted from T. S. Chung, The recent developments of thermotropic liquid crystalline polymers, Polymer Engineering and Science, 26(13), p. 903. Copyright  1986, Society of Plastics Engineers.

types of liquids can occur. A polymer that exhibits crystallinity in the melt and that undergoes an ordered–disordered transformation as a result of thermal effects is called a thermotropic LCP. A polymer requiring a small molecule solvent in order to exhibit crystallinity is termed a lyotropic LCP. All three types of LCP structures can occur in either thermotropic or lyotropic polymers, and both are industrially relevant materials.

98

THE STRUCTURE OF MATERIALS

O

O

C

C

O C

N

N

N

H

H

H

Figure 1.70 The repeat structure of Kevlar .

Perhaps the most widely utilized (and studied) lyotropic LCP is poly p-phenylene terephthalamide (PPTA), more commonly known as Kevlar (see Figure 1.70). Kevlar belongs to the class of aramids that are well known for their LCP properties. Because these polymers are crystalline in solution, they are often spun into filaments, from which the solvent is subsequently removed in order to retain the aligned polymer structure. The result is a highly oriented, strong filament that can be used for a wide variety of structural applications. Most thermotropic LCPs are polyesters or copolymers that can be melted and molded into strong, durable objects.

400

Isotactic polystyrene

Region having no physical meaning

=

Semicrystalline random copolymers Tm

Tg

300

Tg (K)

Tg

=

Tm 74

0.

Isotactic polypropylene Neoprene Natural rubber

200

Silicone rubber

100 200

x

x x

.5T m

Tg

300

m

oly

op

m Ho

ers

x x

x Polyvinyl chloride Nylon 6 x x Nylon 66

Polyvinylidene chloride

Linear polyethylene Branched polyethylene

Block copolymers

=0

400 Tm (K )

500

600

Figure 1.71 The glass transition temperature, Tg , as a function of crystalline melting point, Tm , for homopolymers. Filled circles are addition homopolymers, open circles are elastomers, and crosses are condensation homopolymers. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

STRUCTURE OF COMPOSITES

99

Table 1.26 Predominant Properties of Crystalline Polymers Degree of Crystallinity Temperature Range

Low (5–10%)

Intermediate (20–60%)

Above Tg Below Tg

Rubbery Glassy, brittle

Leathery, tough Hornlike, tough

High (70–90%) Stiff, hard, brittle Stiff, hard, brittle

Rubber Viscous liquid

Mobile liquid

Tm

T °C

Tough plastic

Tg

Crystalline solid 10

1000

Partially crystalline plastic

100,000

10,000,000

Molecular weight

Figure 1.72 Approximate relations among molecular weight, Tg , Tm and polymer properties. From F. W. Billmeyer, Textbook of Polymer Science, 3rd ed. Copyright  1984 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

1.3.7

The Glass Transition

The glass transition is an important parameter, since many physical properties change when polymer chains gain mobility. This is an important temperature because it marks the point where the polymer becomes more easily deformed, or more ductile. As summarized in Table 1.26, totally non crystalline polymers are very brittle below the glass transition, whereas polymers with intermediate crystallinity are very tough above the glass transition. Most homopolymers have glass transition temperatures that are 0.5–0.75Tm , even though the crystalline melting point can vary by several hundred degrees between polymer types (see Figure 1.71). Most plastics are designed to be used between Tm and Tg , as shown in Figure 1.72. To conclude this section on polymols, we should note that we have used the term polymer almost exclusively to refer to organic macromolecules. The term plastic refers not only to organic substances of high molecular weight, but also to such substances that at some point in their manufacture have been shaped by flow. Thus, the term plastic is more specific than the term polymer, and this term carries with it an indication of its processing history. As we will see in Chapter 7, there are many materials that can be considered polymers, yet are formed by routes other than melt processing. 1.4

STRUCTURE OF COMPOSITES

The first three sections of this chapter have described the three traditional primary classifications of materials: metals, ceramics, and polymers. There is an increasing

100

THE STRUCTURE OF MATERIALS

emphasis on combining materials from these different categories, or even different materials within each category, in such a way as to achieve properties and performance that are unique. Such materials are called composites. A composite can be different things, depending on the level of definition we use. In the most basic sense, all materials except elements are composites. For example, a binary mixture of two elements, like an alloy, can be considered a composite structure on an atomic scale. In terms of microstructure, which is a larger scale than the atomic level definition, composites are composed of crystals, phases and compounds. With this definition, steel, which is a suspension of carbon in iron, is a composite, but brass, a single-phase alloy, is not a composite. If we move up one more level on the size scale, we find that there are macrostructural composites: materials composed of fibers, matrices, and particulates—they are materials systems. This highest level of structural classification is the one we will use, so our definition of a composite is this: a material brought about by combining materials differing in composition or form on a macroscale for the purpose of obtaining specific characteristics and properties. 1.4.1

Composite Constituents

The constituents in a composite retain their identity such that they can be physically identified and they exhibit an interface between one another. This concept is graphically summarized in Figure 1.73. The body constituent gives the composite its bulk form, and it is called the matrix. The other component is a structural constituent, sometimes called the reinforcement, which determines the internal structure of the composite. Though the structural component in Figure 1.73 is a fiber, there are other geometries that the structural component can take on, as we will discuss in a subsequent section. The region between the body and structural constituents is called the interphase. It is quite common (even in the technical literature), but incorrect, to use the term interface to describe this region. An interface is a two-dimensional construction—an area having a common boundary between the constituents—whereas an interphase is a three-dimensional phase between the constituents and, as such, has its own properties. It turns out that these interphase properties play a very important role in determining the ultimate properties of the bulk composite. For instance, the interphase is where INTERPHASE (BONDING AGENT) MATRIX FIBER

INTERFACE

Figure 1.73 Schematic illustration of composite constituents. Reprinted, by permission, from M. Schwartz, Composite Materials Handbook, 2nd ed., p. 1.4. Copyright  1992 by McGraw-Hill.

STRUCTURE OF COMPOSITES

101

mechanical stresses are transferred between the matrix and the reinforcement. The interphase is also critical to the long-term stability of a composite. It will be assumed that there is always an interphase present in a composite, even though it may have a thickness of only an atomic dimension. The chemical composition of the composite constituents and the interphase is not limited to any particular material class. There are metal–matrix, ceramic–matrix, and polymer–matrix composites, all of which find industrially relevant applications. Similarly, reinforcements in important commercial composites are made of such materials as steel, E-glass, and Kevlar . Many times a bonding agent is added to the fibers prior to compounding to create an interphase of a specified chemistry. We will describe specific component chemistries in subsequent sections. 1.4.2

Composite Classification

There are many ways to classify composites, including schemes based upon (1) materials combinations, such as metal–matrix, or glass-fiber-reinforced composites; (2) bulkform characteristics, such as laminar composites or matrix composites; (3) distribution of constituents, such as continuous or discontinuous; or (4) function, like structural or electrical composites. Scheme (2) is the most general, so we will utilize it here. We will see that other classification schemes will be useful in later sections of this chapter. As shown in Figure 1.74, there are five general types of composites when categorized by bulk form. Fiber composites consist of fibers; with or without a matrix. By definition, a fiber is a particle longer than 100 µm with a length-to-diameter ratio (aspect ratio) greater than 10:1. Flake composites consist of flakes, with or without a matrix. A flake is a flat, plate-like material. Particulate composites can also have either a matrix or no matrix along with the particulate reinforcement. Particulates are roughly

FIBER COMPOSITE

PARTICULATE COMPOSITE

FLAKE COMPOSITE

LAMINAR COMPOSITE

FILLED COMPOSITE

Figure 1.74 Classes of composites. Reprinted, by permission, from Schwartz, M., Composite Materials Handbook, 2nd ed., p. 1.7. Copyright  1992 by McGraw-Hill.

102

THE STRUCTURE OF MATERIALS

spherical in shape in comparison to fibers or flakes. In a filled composite, the reinforcement, which may be a three-dimensional fibrous or porous structure, is continuous and often considered the primary phase, with a second material added through such processes as chemical vapor infiltration (CVI). Finally, laminar composites are composed of distinct layers. The layers may be of different materials, or the same material with different orientation. There are many variations on these classifications, of course, and we will see that the components in fiber, flake, and particulate composites need not be distributed uniformly and may even be arranged with specific orientations. Cooperative Learning Exercise 1.7 We will see in Section 5.4.2 that the elastic modulus of a unidirectional, continuous-fiberreinforced composite depends on whether the composite is tested along the direction of fiber orientation (parallel) or normal to the fiber direction (transverse). In fact, the elastic modulus parallel to the fibers, E1 , is given by Eq. (1.62), whereas the transverse modulus, E2 , is given by Eq. (1.63). Consider a composite material that consists of 40% (by volume) continuous, uniaxially aligned, glass fibers (Ef = 76 GPa) in a polyester matrix (Em = 3 GPa). Person 1: Calculate the elastic modulus parallel to the fiber direction, E1 . Person 2: Calculate the elastic modulus transverse to the fiber direction, E2 . Compare your answers. What would you expect the composite modulus to be if the same volume fraction of fibers were randomly oriented instead of uniaxially aligned?

Answer : E1 = (0.4)76 + (0.6)(3) = 32.2 GPa; E2 = (76)(3)/[(0.6)(76) + (0.4)(3)] = 4.87 GPa

1.4.3

Combination Effects in Composites

There are three ways that a composite can offer improved properties over the individual components, collectively called combination effects. A summation effect arises when the contribution of each constituent is independent of the others. For example, the density of a composite is, to a first approximation, simply the weighted average of the densities of its constituents. The density of each component is independent of the other components. Elastic modulus is also a summation effect, with the upper limit Ec (u) given by Ec (u) = Vp Ep + Vm Em (1.62) and the lower limit, Ec (l) given by Ec (l) =

Em Ep Vm E p + Vp E m

(1.63)

where Em and Ep are the elastic moduli of the matrix and particulate, respectively; and Vm and Vp are their respective volume fractions. We will discuss mechanical properties in more detail in Chapter 5, but the point of Eqs. (1.62) and (1.63) is that summation properties can be added appropriately to give an estimate of the composite properties. A complementation effect occurs when each constituent contributes separate properties. For example, laminar composites are sandwich-type composites composed of several layers of materials. Sometimes the outer layer is simply a protective coating,

STRUCTURE OF COMPOSITES

103

such as a polymeric film, that imparts corrosion resistance to the composite. This outer layer serves no structural purpose, and contributes only a specific property to the overall composite—in this case, corrosion resistance. Finally, some constituent properties are not independent of each other, and an interaction effect occurs. In this case, the composite property may be higher than either of the components, and the effect may be synergistic rather than additive. For example, it has been observed that the strength of some glass-fiber-reinforced plastic composites is greater than either the matrix or the reinforcement component by itself. In the subsequent sections, we look at the individual components of a composite and see what each can contribute, and how it helps to improved properties of the composite. 1.4.4

The Composite Matrix

The matrix serves two primary functions: to hold the fibrous phase in place and to deform and distribute the stress under load to the reinforcement phase. In most cases, the matrix material for a fiber composite has an elongation at break greater than the fiber; that is, it must deform more before breaking. It is also beneficial to have a matrix that encapsulates the reinforcement phase without excessive shrinkage during processing. A secondary function of the matrix is to protect the surface of the reinforcement. Many reinforcements tend to be brittle, and the matrix protects them from abrasion and scratching, which can degrade their mechanical properties. The matrix can also protect the reinforcement component from oxidation or corrosion. In this way, many fibers with excellent mechanical properties, such as graphite fibers, can be used in oxidizing environments at elevated temperatures due to protection by the matrix constituent. Most fiber-matrix composites (FMCs) are named according to the type of matrix involved. Metal-matrix composites (MMCs), ceramic-matrix composites (CMCs), and polymer-matrix composites (PMCs) have completely different structures and completely different applications. Oftentimes the temperature at which the composite must operate dictates which type of matrix material is to be used. The maximum operating temperatures of the three types of FMCs are listed in Table 1.27. Most structural PMCs consist of a relatively soft matrix, such as a thermosetting plastic of polyester, phenolic, or epoxy, sometimes referred to as resin-matrix composites. Some typical polymers used as matrices in PMCs are listed in Table 1.28. The list of metals used in MMCs is much shorter. Aluminum, magnesium, titanium, and iron- and nickel-based alloys are the most common (see Table 1.29). These metals are typically utilized due to their combination of low density and good mechanical properties. Matrix materials for CMCs generally fall into four categories: glass ceramics like lithium aluminosilicate; oxide ceramics like aluminum oxide (alumina) and mullite; nitride ceramics such as silicon nitride; and carbide ceramics such as silicon carbide. Table 1.27 Approximate Upper Temperature Limits for Continuous Operation of Composites

Composite Polymer matrix Metal matrix Ceramic matrix

Maximum Operating Temperature (◦ C) 400 1000 1300

104

THE STRUCTURE OF MATERIALS

Table 1.28 Some Typical Polymers Used as Matrices in PMCs

Process Temperature (◦ C)

Matrix Material

Upper Use Temperature (◦ C)

Thermosetting Polyester (phthalic/maleic type) Vinyl ester Epoxy Epoxy Phenolic Cyanates (triazines) Bismaleimides Nadic end-capped polyimides (e.g., PMR-15)

RT RT 150 200 250 250 250 316

70 125 125 175 200 200 225 316

325 250 280 280 300 370 350 325 400 370 400 300

180 100 100 100 150 175 175 200 275 316 300 150

Thermoplastic Polysulfone Polyamide Polycarbonate Polyphenylene oxide (PPO) Polysulfides (PPS) Polyether ether ketone (PEEK) Polyether sulfone (PES) Polyamide-imides Polyetherimide Polyimide Polyarylate Polyester (liquid crystalline)

Table 1.29 Common Metals for Metal-Matrix Composites

Matrix Material

Fabrication Method

Aluminum

Diffusion bonding Hot molding Powder metallurgy Liquid processing Liquid processing Diffusion bonding Diffusion bonding Diffusion bonding

Magnesium Titanium Iron-, nickel-based alloys

Typical Composite Density (g/cm3 )

Use Temperature (◦ C) 350

2.62–3.45

1.82–2.80

300

3.76–4.00 5.41–11.7

650 800–1150

The processing techniques used for CMCs can be quite exotic (and expensive), such as chemical vapor infiltration (CVI), or through pyrolysis of polymeric precursors. Their maximum use temperatures are theoretically much higher than most MMCs or PMCs, exceeding 1800◦ C, although the practical use temperature is often much lower

STRUCTURE OF COMPOSITES

105

due to creep deformation (see Table 1.30). The attraction of CMCs, then, is for use in high-temperature structural applications, such as in combustion engines. 1.4.5

The Composite Reinforcement

In Section 1.4.2, we described several classification schemes for composites, including one that is based upon the distribution of the constituents. For reinforced composites, this scheme is quite useful, as shown in Figure 1.75. In reinforced composites, the reinforcement is the structural constituent and determines the internal structure of the composite. The reinforcement may take on the form of particulates, flakes, lamina, or fibers or may be generally referred to as “filler.” Fibers are the most common type of reinforcement, resulting in fiber-matrix composites (FMCs). Let us examine some of these reinforcement constituents in more detail. 1.4.5.1 Fiber-Matrix Composites. As shown in Figure 1.75, there are two main classifications of FMCs: those with continuous fiber reinforcement and those with discontinuous fiber reinforcement. Continuous-fiber-reinforced composites are made from fiber rovings (bundles of twisted filaments) that have been woven into twodimensional sheets resembling a cloth fabric. These sheets can be cut and formed to a desired shape, or preform, that is then incorporated into a composite matrix, typically a thermosetting resin such as epoxy. Metallic, ceramic, and polymeric fibers of specific compositions can all be produced in continuous fashions, and the properties of the Table 1.30 Common Ceramics for Ceramic-Matrix Composites

Matrix Material

Density (g/cm3 )

Use Temperature (◦ C)

Alumina, Al2 O3 Glass ceramics Si3 N4 SiC

4.0 2.7 3.1 3.2

∼1000 900 ∼1300 ∼1300

Composites

Particle-reinforced

Largeparticle

Fiber-reinforced

Dispersion- Continuous strengthened (aligned)

Structural

Discontinuous Laminates (short)

Aligned

Randomly oriented

Figure 1.75 Types of reinforced composites.

Sandwich panels

106

THE STRUCTURE OF MATERIALS

resulting composite are highly dependent not only on the type of fiber and matrix used, but also on the processing techniques with which they are formed. We will discuss continuous-fiber-reinforced composites in more detail in Sections 5.4 and 7.4. Discontinuous-fiber-reinforced composites are much more widely used, and there are some general underlying principles that affect their overall properties. Four main factors contribute to the performance level of a fiber in discontinuously reinforced FMCs. The first factor is fiber orientation. As shown in Figure 1.76, there are several ways that short fibers can be oriented within the matrix. One-dimensional reinforcement occurs when the fibers are oriented along their primary axis. This offers maximum mechanical strength along the orientation axis, but results in anisotropic composites; that is, the mechanical and physical properties are not the same in all directions. Planar reinforcement occurs with two-dimensional orienting of the fibers, as often occurs with woven fabrics. The fabric, as is common in woven glass fibers, is produced in sheets, and it is laid down (much like a laminate) to produce a two-dimensional reinforcement structure. Three-dimensional reinforcement results from the random orientation of the fibers. This creates an isotropic composite, in which the properties are the same in all directions, but the reinforcing value is often decreased compared to the aligned fibers. The second factor that affects performance in discontinuously reinforced FMCs is fiber length. This has an effect primarily on the ease with which the composite can be manufactured. Very long fibers can create difficulties with methods used to create discontinuously reinforced FMCs and can result in nonuniform mechanical properties. The third factor is also related to fiber geometry, namely, the fiber shape. Recall that the

(a)

(b)

(c)

Figure 1.76 Types of fiber reinforcement orientation (a) one-dimensional, (b) two-dimensional, and (c) three-dimensional.

STRUCTURE OF COMPOSITES

107

definition of a fiber is a particulate with length greater than 100 µm and an aspect ratio greater than 10:1. This definition allows for a great deal of flexibility in the geometry of the fiber. For example, aspect ratio can vary widely; many reinforcement filaments have aspect ratios much larger than 10:1. It is also not necessary that the fiber cross section be exactly circular. Hexagonal, ellipsoidal, and even annular (hollow fiber) cross sections are quite common. Finally, it is also not necessary that the fiber, even those with circular cross sections, be exactly cylindrical. “Dumbbell”-shaped fibers are very common, and even preferable, since mechanical stresses tend to concentrate at the fiber ends. The dumbbell shape helps distribute these stresses. The final factor affecting the reinforcement performance is its composition. Chemistry affects properties, and strength is usually the most important property of a reinforcing fiber. Though we will concentrate on mechanical properties of materials in Chapter 5, it is useful at this point to familiarize yourself with some of the common fiber reinforcements, as summarized in Table 1.31. It is worth noting that oftentimes the important design consideration for reinforcement materials (or the matrix, for that matter) is not the absolute value of a particular design criterion, such as tensile strength

Table 1.31 Selected Properties of Some Common Reinforcing Fibers

Fiber

Density (g/cm3 )

Melting Point (◦ C)

Specific Modulus (GPa · cm3 /g)

Specific Strength (MPa · cm3 /g)

Aluminum Steel Tantalum Titanium Tungsten Boron Beryllium Molybdenum Aluminum oxide Aluminosilicate Asbestos Beryllium carbide Beryllium oxide Carbon Graphite E-glass S-glass Quartz SiC (on tungsten) Si3 N4 BN ZrO2 Wood Polyamide (Kevlar) Polyester Polypropylene

2.70 7.87 16.6 4.72 19.3 2.30 1.84 10.2 3.97 3.90 2.50 2.44 3.03 1.76 1.50 2.54 2.49 2.20 3.21 2.50 1.91 4.84 0.4–0.8 1.14 1.40 0.9

660 1621 2996 1668 3410 2100 1284 2610 2082 1816 1521 2093 2566 3700 3650 1316 1650 1927 2316 1900 3000 (decomp.) 2760 — 249 249 154

27 25 12 24 21 192 165 35 132 26 60 127 116 114 230 28 34 32 140 100 47 71 17 2.5 2.9 1.8

230 530 37 410 220 1500 710 140 170 1060 550 420 170 1570 1840 1360 1940 407 1000 1344 722 427 — 730 490 77

108

THE STRUCTURE OF MATERIALS

or modulus, but the value per unit weight, such as specific strength or specific modulus. These values are listed in Table 1.31 rather than absolute values to illustrate this point. This fact is extremely important for many applications, such as automotive and aerospace composites, for which weight savings is paramount. Note also that reinforcing fibers come from all materials classes. Let us examine some of the more common fibers in more detail. Organic fibers generally have very low specific gravities, so they are attractive for applications where strength/weight ratio is important. They are also very common for textile applications, since some organic fibers are easily woven. Carbon fibers offer excellent thermal shock resistance and a very high strength-to-weight ratio. There are two general types of carbon fibers, PAN-based carbon fibers and pitch-based carbon fibers. PAN-based carbon fibers are manufactured by the pyrolysis of polyacrylonitrile (PAN), as illustrated in Figure 1.77. The PAN is polymerized and fibers spun from the pre-polymer. The PAN fibers are then pyrolyzed to remove the hydrogens and form benzene-ring structures. Pitch-based fibers are produced in a similar pyrolysis process of a precursor fiber, except that the precursor in this case is pitch. Pitch is actually a liquid crystalline material, often called mesophase pitch, and is composed of a complex mixture of thousands of different species of hydrocarbons and heterocyclic molecules. It is the residual product of petroleum refining operations. Wood fibers are technically organic, and though we do not discuss wood as a separate materials class in this text, it is an important structural material. Wood fibers, often in the form of wood flour, possess a variety of properties depending upon the type of tree from which they are derived, but are used extensively in low-cost composites. Wood fiber has a good strength/weight ratio and provides a use for recycled paper products. Turning to synthetic organic fibers, we see that polyamide fibers, such as Kevlar , offer excellent specific mechanical properties. Kevlar is used in many applications where high toughness is required, such as ropes and ballistic cloths. In addition to polyamides, polyesters, and polypropylene fibers listed in Table 1.31, nylon and polyethylene are other common polymeric fibers used for composites. In all cases, an added attraction of synthetic fibers is their chemical inertness in most matrix materials. Glass fibers are the most common reinforcing fiber due to their excellent combination of mechanical properties, dielectric properties, thermal stability and relatively low cost. As a result, there are many different types of silicate glass fibers, all with varying properties designed for various applications (see Table 1.32). The majority

N

N C

N C

C

CH

CH

CH CH2

CH2 N

C

C

C C

CH CH2

N C

C C

Figure 1.77

N C

C

CH CH2

C

N C

C C

Pyrolysis of polyacrylonitrile (PAN) to form carbon fibers.

STRUCTURE OF COMPOSITES

109

Table 1.32 Composition of Commercial Glass Fibers Composition (wt%) SiO2 A-glass (typical) E-glass (range) AR glass (range) C-glass (range) S and R glasses (range)

73

Al2 O3 1

52–56 12–16

Fe2 O3

B2 O3

ZrO2

0.1 0–0.5

8–13

60–70 0–5

15–20

59–64 3.5–5.5 0.1–0.3 6.5–7

MgO

CaO

4

8

0–6

16–25

Na2 O 13

0–10

K2 O

Li2 O TiO2

F2

0.5 ← TA A

T 4 = TA

Xb

B

A

Xb

(a)

B

(b)

O

O b

b a

∆Gmix

L

∆Gmix

L

a

T3 = TB Xb

A

T2 < TB A

B

Xb (d)

(c)

T5 T4

O

L

T3 L a

L +b b a

a+ b

T1 = Te Xa

A

L +a

T2 T T1

b ∆Gmix

B

Xb

XL Xb

B

Xa

A

(e)

Xb

XL Xb

B

(f)

Figure 2.6 Free energy of mixing for solid and liquid phases at various temperatures (a–e) and resulting temperature–composition phase diagram for a mostly insoluble binary component system (f). From O. F. Devereux, Topics in Metallurgical Thermodynamics. Copyright  1983 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc. Table 2.2 Summary of Phase Relationships in a Phase Diagram ž One-phase regions may touch each other only at single points, never along boundary lines. ž Adjacent one-phase regions are separated from each other by two-phase regions involving

the same two phases. ž Three two-phase regions must originate upon every three-phase isotherm; that is, six

boundary lines must radiate from each three phase reaction horizontal. ž Two three-phase isotherms may be connected by a two-phase region provided that there are

two phases which are common to both of the three-phase equilibria.

156

THERMODYNAMICS OF CONDENSED PHASES

respectively, such that Wα + Wβ = 1

(2.38)

Also, the mass of each component, in this case Ni and Cu, distributed between the two phases must equal their masses in the original composition: Wα CS + Wβ CL = C0

(2.39)

Solution of Eqs. (2.38) and (2.39) gives the weight fraction of liquid in terms of the compositions CS − C0 (2.40) Wβ = CS − CL and the weight fraction of solid in terms of compositions Wα =

C0 − CL CS − CL

(2.41)

Equations (2.40) and (2.41) are the lever rule and can be used to determine the relative amounts of each phase in any two-phase region of a binary component phase diagram. For the example under consideration, the amount of liquid present turns out to be Wβ =

63 − 53 = 0.625 63 − 47

(2.42)

So 62.5% by weight of the alloy is present as liquid, and 37.5% is solid. The compositions of the two phases are determined by extending the horizontal line, or tie line, to the phase boundaries (line EF in Figure 2.4), then reading the compositions of these phases off the abscissa. The liquid phase has a composition of 47 wt% Ni/53 wt% Cu (point E), and the solid phase is 63 wt% Ni/37% Cu (point F). There are a few important points regarding application of the lever rule. First, it can only be used in the form of Eqs. (2.40) and (2.41) when concentrations are expressed in terms of weight percentages. Mole percentages must first be converted to weight percents when concentrations are so expressed. Second, Eqs. (2.40) and (2.41) can be used by simply measuring the distances of the corresponding line segments directly off the diagram. For example, the line EF in Figure 2.4 represents the quantity CS − CL . If it is measured using a ruler in, say, millimeters, then CO − CL could be measured similarly, and Wα determined from a ratio of these two measurements. In fact, this method is much more accurate than extrapolating lines from the phase boundaries down to the abscissa (which can introduce error) and then interpolating the composition from markings that may, in some cases, only give markers every 20 wt% or so. Finally, as you begin to work problems with the lever rule, you may find that the ratio of line segments given by Eqs. 2.40 and 2.41 is a little bit counterintuitive. That is, the line segment in the numerator for the weight fraction of solids [Eq. (2.41)] is that opposite the solid-liquid + solid phase boundary, and the line segment in the numerator for the liquid fraction [Eq. (2.42)] is opposite the liquid-solid + liquid phase boundary. A little practice will eliminate this confusion.

THERMODYNAMICS OF METALS AND ALLOYS

157

Cooperative Learning Exercise 2.2 1500

1300

MgNi2 + L

L+b

T (°C)

1100

L L + MgNi2

900

MgNi2 b

700

L + Mg2Ni

a+L 500 a 300

0 (Mg)

Mg2Ni Mg2Ni + MgNi2

a + Mg2Ni 20

MgNi2 + b

40

60 Wt% Ni

80

100 (Ni)

Reprinted, by permission, from K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering , p. 333. Copyright  1976 by John Wiley & Sons, Inc.

Consider the Mg–Ni binary phase diagram shown above. Note that in this system, intermetallic compound formation is common; for example, Mg2 Ni, MgNi2 are stoichiometric compounds. Person 1: Determine the phases present, the relative amounts of each phase, and the composition of each phase when the only incongruently melting intermetallic compound in this system is heated to 800◦ C. Person 2: Determine the phases present, the relative amounts of each phase, and the composition of each phase when the liquid phase from Person 1’s problem is cooled to 600◦ C. Answer : Person 1: 88% L (51 wt% Ni, 49 wt% Mg), 12% MgNi2 (83 wt % Ni, 17 wt% Mg). Person 2: 14% L (27 wt% Ni, 73 wt% Mg); 86% Mg2 Ni (55 wt% Ni, 45 wt % Mg).

2.1.2.3 Three-Phase Transformations in Binary Systems. Although this chapter focuses on the equilibrium between phases in binary component systems, we have already seen that in the case of a eutectic point, phase transformations that occur over minute temperature fluctuations can be represented on phase diagrams as well. These transformations are known as three-phase transformations, because they involve three distinct phases that coexist at the transformation temperature. Their characteristic shapes as they occur in binary component phase diagrams are summarized in Table 2.3. Here, the Greek letters α, β, γ , and so on, designate solid phases, and L designates the liquid phase. Subscripts differentiate between immiscible phases of different compositions. For example, LI and LII are immiscible liquids, and αI and αII are allotropic solid phases (different crystal structures).

158

THERMODYNAMICS OF CONDENSED PHASES

Table 2.3

Common Three-phase Transformations in Condensed Binary Systems

Name of Reaction

Equation

Phase Diagram Characteristic

cooling

Monotectic

LI −−−−→ α + LII

Monotectoid

α1 −−−−→ α2 + β

Eutectic

L −−−−→ α + β

Eutectoid

α −−−−→ β + γ

LI a1

cooling

a2

cooling

L

Peritectic

α + L −−−−→ β

Peritectoid

α + β −−−−→ γ

b a

b

cooling

LI + LII −−−−→ α

b

a

cooling

Syntectic

LII

a

LI

cooling

g LII

a

L

a b

cooling

a

b g

Source: K. M. Ralls, T. H. Courtney and J. Wulff, Introduction to Materials Science and Engineering, p. 331. Copyright  1976 by John Wiley & Sons, Inc. 1100 a+L

L b+L

900

700

T(°C)

+ a

g+L g + d+L d d

b

a

+

b

b

g

g d+L

500

∋ b′

h+L



h

g+L ∋

0 (Cu)

20

b′ + g

40 Wt. % Zn

60

d+h



a + b′

300

80

100 (Zn)

Figure 2.7 Cu–Zn phase diagram, illustrating a number of three-phase reactions. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

THERMODYNAMICS OF METALS AND ALLOYS

159

The Cu–Zn system (see Figure 2.7) displays a number of intermediate solid solutions that arise due to limited solubility between the two elements. For example, at low wt% Zn, which incidently is the composition of alloys known as brass, the relatively pure copper α phase is able to accommodate small amounts of Zn as an impurity in the crystal structure. This is known as a terminal solid phase, and the solubility limit where intermediate solid solutions (such as α + β) begin to occur is called the solvus line. Some of the three-phase transformations that are found in this diagram include a peritectic (δ + L → ε) and a eutectoid (δ → γ + ε). Remember that these three-phase transformations are defined for equilibrium cooling processes, not heating or nonequilibrium conditions. 2.1.3

The Iron–Carbon Phase Diagram*

Iron and its alloys continue to be the most widely utilized metallic systems for construction. The process by which iron-based alloys, including steels, are produced from iron

HISTORICAL HIGHLIGHT

Iron was the first metal for the masses. As far back as five thousand years ago, metalworkers must have somehow gotten their furnaces especially hot on a day when they had charged the furnace with some iron-rich rock, maybe blood-red hematite. Perhaps it was a consistent blast from the sea that made the difference. From this they obtained a silver-gray metal that they had never seen before. It was iron. One of the oldest known iron implements is a dagger blade from 1350 B.C. The downside of iron is that wrenching it from ore requires a temperature hundreds of degrees higher than that needed to extract copper or tin. In fact, the melting temperature was too high for ancient furnaces, so early iron-workers got their metal by beating it out of a solid “bloom,” a conglomeration of metal and rocky slag that forms when the ore is heated to temperatures high enough to loosen things up in the ore but not high enough to melt iron. The earliest iron turned out to be no better than bronze. Often it was worse, especially for weaponry. Compared to copper and bronze, it was a slightly crazy metal, with unpredictably varying qualities, so a smith never quite knew what to expect from a finished iron implement or weapon—often they came out too soft or too brittle.

The reason for this fickleness is the special role that carbon atoms play in the properties of iron-based metals, including steel. No one knew in the early days of iron that carbon atoms from the burning fuel in their bloomery and smithing furnaces were actually becoming part of the metal’s internal anatomy and changing its properties. In the smelting of copper or tin, carbon’s role is to remove oxygen without remaining in the metal. The iron that smiths wrought from the blooms and then formed into weapons and implements absorbed varying amounts of carbon. Too little carbon resulted in a consistency closer to that of pure iron, which is soft and malleable. A man wielding a pure iron sword would be no match for an opponent swinging a bronze sword. Too much carbon in the iron, however, yielded a brittle metal that could shatter like pottery. Most often, smiths ended up with iron containing somewhere around 1 to 3 percent carbon by weight. When lucky, they would get somewhere between .1 and 1 percent carbon. When that happened, the metal underwent a dramatic personality change. It became much harder and tougher, and it could hold a much sharper edge. When a smith managed to keep the carbon content within this range, he made steel. Source: Stuff, I. Amato, pp. 28–29.

160

THERMODYNAMICS OF CONDENSED PHASES

ore involves the reduction of iron oxide by means of carbon to form a product known as pig iron, which contains approximately 4% carbon. As a result, the iron–carbon phase diagram is an industrially important one. The compounds formed between iron and carbon are sufficiently complex in their microstructures and properties to warrant intensive investigation and more detailed description. However, the iron–carbon phase diagram is really just another binary phase diagram, and, as such, we have all the tools we need to fully understand it. The iron–carbon phase diagram at low weight percentages of carbon is shown in Figure 2.8. Actually, the phase boundary at 6.69 wt% carbon represents the compound iron carbide, Fe3 C, known as cementite, so that the phase diagram in Figure 2.8 is more appropriately that of Fe3 C–C. The allotropy of elemental iron plays an important role in the formation of iron alloys. Upon solidification from the melt, iron undergoes two allotropic transformations (see Figure 2.9). At 1539◦ C, iron assumes a BCC structure, called delta-iron (δ-Fe). Upon further cooling, this structure transforms to the FCC structure at 1400◦ C, resulting in gamma-iron (γ -Fe). The FCC structure is stable down to 910◦ C, where it transforms back into a low-temperature BCC structure, alpha-iron (α-Fe). Thus, δ-Fe and α-Fe are actually the same form of iron, but are treated as distinct forms due to their two different temperature ranges of stability. Carbon is soluble to varying degrees in each of these allotropic forms of iron. The solid solutions of carbon in α-Fe, γ -Fe, and δ-Fe are called, respectively, ferrite, austenite, and δ-ferrite. So, for example, the single-phase region labeled as γ in

1700

d+L 0.09 0.17 0.53 1495°C

1538 1500 d 1394

L

d+g

1300

g+L

L + Fe3C

1148°C

T(°C)

1100

2.11

g

912 900

4.30

g + Fe3C

a d+g 700 0.02 0.80

Fe3C

723°C a + Fe3C

500

0 (Fe)

1

2

3 Wt. % C

4

5

6

6.69 (Fe3C)

Figure 2.8 The Fe–C phase diagram (low weight % C). From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

THERMODYNAMICS OF METALS AND ALLOYS

1539° 1400°

Liquid d

d-Iron, BCC

161

Solid, d g

T (°C)

g-Iron, FCC

910°

g

a

a-Iron, BCC non-magnetic 768°

a-Iron, BCC magnetic

Time

Figure 2.9 Cooling curve for pure iron. Reprinted, by permission, from Committee on Metallurgy, Engineering Metallurgy, p. 245. Copyright  1957 by Pitman Publishing.

Figure 2.8 could be (and is in many diagrams) labeled as the austenite phase region. Similarly, the α and δ phase can be called the ferrite and δ-ferrite regions. Keep in mind, however, that these single-phase regions are simply a solid solution between one of the allotropic forms of elemental iron and carbon. Note that the ferrite region is particularly limited in both temperature and composition, as is δ-ferrite, and that austenite has a particularly large range of stability at higher temperatures. All alloys with carbon contents between 0.5 and 2 wt% C solidify as austenite. Alloys containing more than 2% carbon are subject to a eutectic transformation that forms austenite and cementite. At 0.8% carbon, what has solidified as austenite undergoes a eutectoid transformation at 723◦ C to an intimate mixture of ferrite and cementite. This mixture is given a special name, pearlite. Recall that phases need not be continuous, and such is the case in pearlite where the cementite (Fe3 C) forms lamellar structures within the austenite phase, which can also have some carbon dissolved in it (see Figure 2.10). The eutectoid transformation is an important one not only for this specific carbon composition, but for classifying all types of steels. Carbon steels have carbon contents between 0.1 and 1.5 wt%. Those with carbon contents less than 0.8% are termed hypoeutectoid steels, and those with greater than 0.8% C are called hypereutectoid steels. Further classifications of steels are given in Table 2.4. When austenite is cooled under more rapid conditions, a compound called bainite is produced. Bainite is a nonequilibrium product that is similar to pearlite, but consists of a dispersion of very small Fe3 C particles between the ferrite plates. Bainite formation is favored at a high degree of supercooling from the austenite phase, whereas pearlite forms at low degrees of supercooling, or more equilibrium cooling. Finally, if an austenite steel is rapidly quenched, martensite steel can form. Steel martensite possesses a body-centered tetragonal structure (see Figure 2.11), which is

162

THERMODYNAMICS OF CONDENSED PHASES

1100

1000

g

g + Fe3C

900 g

Temperature (°C)

x

g

800

a a+g

g g

727°C

b

700

a

a 600

Fe3C

a + Fe3C

500

400

x′ 0

1.0

2.0

Composition (wt % C)

Figure 2.10 Schematic representation of the microstructures for a eutectoid transformation in the Fe–C system. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 277. Copyright  2000 by John Wiley & Sons, Inc.

Table 2.4

Classification of Ferrous Alloys

Component (wt%)

Classification

C < 0.01 0.1 < C < 1.5

Iron Carbon steel

2 < C < 4.5

Cast iron

Various

Alloy steels

Cr > 12

Stainless steels

Examples

1000 series—ferrite/pearlite 2000 series—ferrite/pearlite or bainite 3000 series—martensitic 4000 series Gray iron White iron Ni, Mn, Cu, Co, C, N—austenite stabilizers Cr, Mo, Si, W, V, Sn, Cb, Ti—ferrite stabilizers 200/300—austenitic 400—ferritic 400/500—martensitic

THERMODYNAMICS OF METALS AND ALLOYS

163

= Fe atoms = C atom sites

c

a a

Figure 2.11 The body-centered tetragonal unit cell of steel martensite. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

0.305

Lattice parameter (nm)

0.300 0.295

c parameter

0.290

a parameter

0.285 0.280

0

0.2

0.4

0.6 0.8 Wt.% C

1.0

1.2

1.4

Figure 2.12 Variation of tetragonality in steel martensite with axial ratio. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

only slightly distorted from the BCC structure of austenite. The distortion arises from interstitial carbon atoms that occupy one set of preferred octahedral interstitial sites. During the martensitic transformation upon rapid cooling, carbon atoms residing at equivalent octahedral sites in BCC change position slightly to only one (along the c-axis) of the three possible sets of octahedral sites in the body-centered tetragonal structure, thus creating the distortion. The extent of tetragonal distortion, as measured by the axial ratio, increases with increasing carbon content (see Figure 2.12). The transformation process from austenite to martensite is very rapid, because it involves only a distortion of the BCC unit cell, as opposed to the BCC–FCC transition to pearlite or bainite, which is a diffusion-controlled process. This type of diffusionless transformation is termed a martensitic transformation and can occur in alloys other than steel.

164

THERMODYNAMICS OF CONDENSED PHASES

I

Temperature

II

Zone III

Commercial cast iron range 3.0 0.8 Wt.% C

I II III

4.3

Fe3C Desulfurized

Fast cool γ+L

Moderate γ+L

Slow cool γ+L

I

γ + Fe3C P + Fe3C

γ + Gf P + Gf

γ + Gf

II

γ + Gn

γ + Gn

α + Gf

III

P + Gn

α + Gn

White cast iron

Pearlitic gray cast iron Reheat; hold in zone II 30+ hours Fast cool

Slow cool

II

γ + Gr

γ + Gr

III

P + Gr

α + Gr

Pearlitic malleable cast iron

Ferritic gray cast iron

Moderate γ+L

C

Pearlitic ductile cast iron

Slow cool γ+L

Ferritic ductile cast iron

Ferritic malleable cast iron

Figure 2.13 Microstructures obtained by varying thermal treatments in cast irons (Gf = graphite flakes; Gr = graphite rosettes; Gn = graphite nodules; P = pearlite). From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

THERMODYNAMICS OF CERAMICS AND GLASSES

165

As you can see, the process by which the iron–carbon alloy is processed and solidified is just as important as the overall stoichiometry. Although a discussion regarding phase transformations is more the realm of kinetic processes, it is nonetheless pertinent to summarize here the types of important ferrous alloys, particularly those in the cast iron categories. This is done in Figure 2.13. As indicated in Table 2.4, cast irons contain more than 2% carbon by weight, compared to steels which usually have less than 1.5% carbon. Cast irons also differ microstructurally from steels in that a separate graphite phase typically exists. Graphite is seldom found in steels because the solid-state transformations that give rise to it are so slow that transformations yielding cementite always predominate. In cast irons, however, graphite can form directly from eutectic solidification and by solid-state transformations. Whether carbon is present as graphite or is tied up in cementite in cast irons depends on whether the solidification and cooling processes are carried out under close to equilibrium conditions, which favor graphite formation, or under highly nonequilibrium conditions, which favor Fe3 C formation. There is an analogy here with austenite decomposition that we have already described: Pearlite forms under near-equilibrium cooling conditions in preference to the less stable bainite, which forms under rapid cooling conditions. As shown in Figure 2.13, slow cooling rates favor graphite formation from the eutectic reaction. The graphite phase continues to grow with decreasing temperature as the carbon content of the iron-rich phase (γ or α) decreases. Under these conditions, the phases present in zones I, II, and III are austenite and liquid, austenite and graphite, and ferrite and graphite, respectively. Under normal conditions, the graphite is present as flakes, but if the melt is desulfurized (where sulfur is an impurity in the ore), the graphite will be present as nodules. The normal structure is called ferritic gray cast iron, and that for the desulfurized material is called ferritic ductile cast iron. Under moderate cooling rates, graphite still forms in the eutectic reaction, but Fe3 C is one product of the eutectoid reaction. The final structure, consisting of locally interconnected graphite flakes dispersed in a pearlitic matrix, is called pearlitic gray cast iron or, if desulfurized, pearlitic ductile cast iron. White cast irons are formed under rapid cooling, and have similar structures to steel insofar as there is no graphite present. Thus, austenite and Fe3 C are present in zone II, and pearlite and Fe3 C are present in zone III. White casts can be converted into more usable structures by heating into zone II for fairly long time periods. This treatment causes graphite, in the shape of rosettes, to precipitate and the cementite is eliminated. Keep in mind that the cooling classifications of “slow,” “medium,” and “rapid,” are arbitrary and that commercial processes can produce materials with a mixture of the various microstructures shown in Figure 2.13. We will discuss transformation processes and the effect of heating and cooling rates on them in Chapter 3.

2.2 2.2.1

THERMODYNAMICS OF CERAMICS AND GLASSES Phase Equilibria in Ternary Component Systems

Three-component systems, or ternary systems, are fundamentally no different from two-component systems in terms of their thermodynamics. Phases in equilibrium must still meet the equilibrium criteria [Eqs. (2.14)–(2.16)], except that there may now be as many as five coexisting phases in equilibrium with each other. The phase rule still

166

THERMODYNAMICS OF CONDENSED PHASES

applies to ternary systems, and we will see that we can use an adaptation of the lever rule to three-component phase diagrams. As with unary and binary component systems, the three components of a ternary system can be anything, even elements, as in our preceding descriptions. In fact, there are many important alloy systems that consist of three primary components, such as iron–carbon–silicon and aluminum–zinc–magnesium. However, the addition of one important element, oxygen, to any metallic binary component system suddenly transforms the discussion from that which is based primarily in metallurgy, to one which is more the realm of ceramics. For example, the aluminum–silicon binary metal alloy system becomes a ternary system of aluminosilicates, and an important class of ceramics, when oxygen is introduced as the third component, and compounds such as Al2 O3 and SiO2 can form. Thus, the discussion here will be limited to ceramics, but keep in mind that the principles of ternary phase diagrams can be applied to any three-component system, including three compounds. Again, let us consider only condensed phases. 2.2.1.1 Ternary Phase Diagrams. In a ternary system, it is necessary to specify temperature, pressure, and two composition parameters to completely describe the system. Typically, pressure is fixed, so that there are three independent variables that are needed to fix the system: temperature and two compositions. The third composition is, of course, fixed by the first two. We could create a three-dimensional plot with three mutually perpendicular axes, as is usually the case in mathematics; however, it is more convenient, and graphically more appealing, to establish two compositional axes 60◦ apart from each other, with a third, redundant compositional axis, as in the form of an equilateral triangle (see Figure 2.14). The temperature axis is then constructed perpendicular to the plane of the triangle, if desired. Figure 2.14 requires a bit of instruction and practice before it can be used correctly. First of all, the pure components—in this case A, B, and C—are located at the vertices

Pure A

0 c′ 0.2

0.8

0.4

0.6

Xa

Xc 0.6

0.4

X′ 0.8

0.2

a′ 0 Pure B

0.8

0.6

0.4

Xb

b′

0.2

0 Pure C

Figure 2.14 Illustration of how to express compositions on a three-component diagram. From O. F. Devereux, Topics in Metallurgical Thermodynamics. Copyright  1983 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

THERMODYNAMICS OF CERAMICS AND GLASSES

167

of the equilateral triangle. The base of the triangle opposite the corresponding vertex refers to 0% of that component, so that the edge of the triangle connecting vertices B and C is 0 wt% A. Lines parallel to each edge constitute lines of constant composition for one of the components. For example, lines parallel to BC, which are spaced a uniform distance apart and become shorter in length as they approach the vertex A, are lines of constant wt% A. Sometimes these lines are every 10 wt%, sometimes every 20 wt%, and sometimes they are not there at all. Similarly, lines parallel to AC and AB represent lines of constant composition for components B and C, respectively. The triangular grid generated by these lines of constant composition helps us determine the overall composition of a point in the diagram. For example, point X′ in Figure 2.14 is located at 30% A, 50% B, and 20% C. Obviously, the overall weight percentages of the components must add up to 100%. Not all compositions fall neatly on the intersection of three grid lines, of course, but when the grid is present, it is relatively easy to interpolate between gridlines. Gridlines can clutter the phase diagram, however, especially when phases and boundaries start appearing in real systems, so that it is convenient to have a method for determining the composition of any point on the diagram without the assistance of gridlines. This is accomplished using the center of gravity rule, which states that the sum of the perpendicular distances from any point within an equilateral triangle to the three sides is constant and equal to the altitude of the triangle. With reference to Figure 2.15, this means that the quantity (a + b + c) is a constant, regardless of where the point is located inside the triangle, and it represents the length from any edge to its opposite vertex.∗ Moreover, the relative amounts of the components are then given

A

80

%

B

20 40

60

c

40

a

80

B

20

40

60

A

%

b

60

20

80

C

%C

Figure 2.15 Illustration of center-of-gravity rule for determining compositions in ternary system. ∗

Don’t let the different ways the axes are labeled between Figures 2.14 and 2.15 confuse you. For example, the labels for the gridlines of component. A are on the left in Figure 2.14, but on the right in Figure 2.15. Similar adjustments are made to the labels for components B and C. Both methods and both figures are entirely correct. You should convince yourself of this fact.

168

THERMODYNAMICS OF CONDENSED PHASES

by the ratio of their respective line segments to the altitude: 100a (a + b + c) 100b %B = (a + b + c) 100c %C = (a + b + c) %A =

(2.43a) (2.43b) (2.43c)

Just as with the lever rule in binary phase diagrams, segments can be measured in any consistent units of length.

Cooperative Learning Exercise 2.3

Using the diagram in Figure 2.15: Person 1: Lightly sketch in grid lines on the diagram and use them to determine the composition at the indicated point. Person 2: Use the center-of-gravity equations (2.43) to calculate the composition at the indicated point. Compare your answers. Answer : 37% A; 21% B; 42% C

Now that we can determine overall compositions on a ternary diagram, we can return to the issue of representing temperature in the diagram. As mentioned earlier, the temperature axis is perpendicular to the compositional, triangular plane, as shown in Figure 2.16 for the case of three components that form a solid solution, α. Recall that in order to form solid solutions, the components must generally have the same crystal structure. Such substances are said to be isomorphous. There are several important things to note in Figure 2.16. First, we are assuming a fixed pressure, as is the case in most condensed-phase diagrams. Second, the phase rule still applies. The single-phase regions, α or liquid, have three degrees of freedom, as given by the phase rule F =C−φ+N =3−1+1=3

(2.44)

so that composition of all three components can be varied continuously in these regions. In Figure 2.16, there is only one two-phase region, L + α. Here, there are only F = 3 − 2 + 1 = 2 degrees of freedom. The upper surface of the two-phase region is the liquidus, and the lower surface of the region is the solidus. These phase boundary surfaces are indicated by selected solid and dashed lines, respectively. Finally, you should be able to visualize that each vertical face of the diagram represents a complete binary component phase diagram. For example, the C–B face in Figure 2.16 occurs only at A = 0 for all temperatures, and the resulting diagram is one similar to the binary component phase diagram for isomorphous substances shown in Figure 2.4. Similarly, the faces represented by A–B (C = 0) and A–C (B = 0) are independent

THERMODYNAMICS OF CERAMICS AND GLASSES

169

L

Temperature

L+α

Solidus

Temperature

Liquidus

α

A

B

po

m

Co

n

itio

sit

pos

n

io

m Co

C

Figure 2.16 Temperature–composition space diagram of a ternary isomorphous system. Reprinted, by permission, from F. N. Rhines, Phase Diagrams in Metallurgy. Copyright  1956 by McGraw-Hill Book Co.

binary phase diagrams. Thus, a ternary phase diagram contains not only all of the phase information we will be describing in this section, but three independent binary component phase diagrams as well. Before moving on to more complicated ternary phase diagrams, it should be pointed out that the complete temperature–composition ternary component phase diagram can be “sectioned” in two important ways in order to simplify the representation. First, we can take horizontal “slices” of the diagram. Such sections are parallel to the base of the diagram and are, as such, necessarily at a fixed temperature, or isotherm. Examples of isothermal sections from a temperature–composition ternary phase diagram are shown in Figure 2.17. Note that at temperatures above the melting point of the highest-melting compound, the resulting diagram consists completely of a single-phase liquid region. As we decrease temperature and take a section below the liquidus, as in Figure 2.17a, part of the diagram contains liquid, part contains solid, and part contains the liquid + solid two-phase region. The liquidus and solidus lines can now be identified, and what was once a two-phase region (L + α) with three degrees of freedom in Figure 2.16 is now a two-phase region with two degrees of freedom, just as in a binary-component diagram, since the temperature has now been fixed and one degree of freedom has been lost. The two degrees of freedom in this region are both compositions, with the third compositional variable being fixed when the other two are determined. Unlike true binary-component phase diagrams, the tie lines connecting the two conjugate phases (in this case liquid and α) that coexist at temperature T1 cannot be drawn as completely

170

THERMODYNAMICS OF CONDENSED PHASES

T1 T2 T3

A

B

Conjugate phases

C A

L

α B +α

B

B

C

C

A

Solidus

A

B

α Solid

us

A

B α

s

u id qu Li

L

A

L+α

T1 XA = XA′

L

C

C

(a)

(b)

T3

T2 Liquidus

C (c )

Figure 2.17 Isotherms through the ternary isomorphous phase diagram. Reprinted, by permission, from F.N. Rhines, Phase Diagrams in Metallurgy. Copyright  1956 by McGraw-Hill Book Co.

horizontal or vertical lines, but must be determined experimentally. These tie lines are represented as dashed lines in the two-phase region. A similar type of diagram results from an isothermal section taken at T2 , as in Figure 2.17b, which is slightly lower than T1 . Finally, at T3 , the isotherm is completely within the single-phase solid region. The second type of “slice” through the temperature-composition is a vertical section, known as an isopleth. Isopleths are sections of constant relative composition. For example, in Figure 2.18a, the line X–B is drawn on the base of the diagram and extended upwards along the temperature axis to generate the diagram shown below, known as a pseudobinary phase diagram, since it looks similar to a binary-component phase diagram, but is obviously different. Note, for example, that the liquidus and solidus lines do not converge at X on the pseudobinary phase diagram. This section has been selected such that A and C are always in the ratio of 1:1 at any point on the diagram. Obviously, the absolute amount of B can change, from 0% at X to 100% at B, thus changing the absolute amounts of A and C, but A and C always stay in the same proportion relative to one another. Vertical sections need not bisect a vertex. They are, however, normally taken in a logical manner—for example, a constant composition of one component or constant relative composition of two components. Figure 2.18b shows a section taken at a constant composition for A, as indicated by the line Y–Z, extended upward along the temperature axis. The resulting pseudobinary diagram is for a fixed composition of A, where the relative amounts of C and B are allowed to vary between pure C at Y and pure B at Z.

THERMODYNAMICS OF CERAMICS AND GLASSES

z

B A

A

171

B

x Y C

C L

s uidu

Liq

L+α

L

s

lidu

So

dus iqui

L

L+α

α

X

α

B (a)

s

lidu

So

Y

Z ( b)

Figure 2.18 Examples of isopleths and pseudobinary phase diagrams. (a) constant A/C; (b) constant A. Reprinted, by permission, from F. N. Rhines, Phase Diagrams in Metallurgy. Copyright  1956 by McGraw-Hill Book Co.

Just as in the case of binary systems, ternary systems can contain components that have only partial solubility on one another. This leads to eutectic diagrams, such as the one shown in Figure 2.19. This is the three-component analogue to Figure 2.5. Again, each face is a binary phase diagram, with binary eutectic melting points located at e1 , e2 , and e3 . The ternary eutectic melting point is located at point E, which is a “well” in the middle of the diagram. The lines connecting the binary eutectics with the ternary eutectics, e1 –E, e2 –E, and e3 –E, are part of the liquidus surface, separating the single-phase liquid region from the solid solution regions. There are, of course, numerous combinations of solid solutions, phases, and transformations possible in these diagrams, and we will not attempt to address more than a few of the basic ones here. But keep in mind that these phase diagrams are generated in the same way that the unary and binary phase diagrams were generated: from free energy versus composition curves at various temperatures. The calculation of free energies of mixing for three components is directly analogous to that for two components, and the analogues for the minimization of free energy criteria as given by Eqs. (2.37) are appropriate.

172

THERMODYNAMICS OF CONDENSED PHASES

Cooperative Learning Exercise 2.4 Work in groups or three. The melting points of components A, B, and C are 1000◦ C, 900◦ C, and 750◦ C, respectively. The constituent binary systems of the ternary system ABC each show complete solubility in both the liquid and the solid states. The following data refer to three binary alloys:

Composition (wt%) A B C 50 50 —

50 — 50

— 50 50

Temperature, ◦ C Liquidus Solidus 975 920 840

950 850 800

Person 1: Sketch and label the AB binary temperature–composition phase diagram. Person 2: Sketch and label the BC binary temperature–composition phase diagram. Person 3: Sketch and label the AC binary temperature–composition phase diagram. Combine your information to sketch the ABC temperature–composition ternary phase diagram, assuming that the liquidus and solidus surfaces do not exhibit a maximum or minimum. Now take a vertical section joining 100% A to the mid-point of the BC binary face. Sketch the resulting pseudo-binary phase diagram. Answers: Each binary phase diagram will look like Figure 2.4, with the melting point of each component the appropriate end of each diagram. The ternary phase diagram is shown above, along with the vertical section joining the vertex at A to the 50% BC composition. C A

B

A Temperature (°C)

800 900 1000

50B :50C

750 °C

S L+

S

L

900 °C 1000 °C

2.2.1.2 Ceramic Phase Diagrams. Let us now examine a real three-component phase diagram. This could be three elements—for example, copper—with any of a number of other metallic elements that lead to an important class of alloys called bronzes. Copper–zinc–manganese alloys are manganese bronzes, copper–tin–nickel alloys are nickel–tin bronzes, copper–tin–lead alloys are leaded tin bronzes, and so on. But we are not limited only to metals in ternary phase diagrams. When we introduce nonmetallic elements, such as carbon, nitrogen, and oxygen, we enter a new realm of materials engineering and science—that of ceramics. The compounds formed in

THERMODYNAMICS OF CERAMICS AND GLASSES

173

C

A e2 a1

a2

c1

B

c2

β

e3

e1 a

c b1 E e2′

A′

e1

C′

E

b e3′

B′

Figure 2.19 Perspective drawing of a ternary system with a simple eutectic and no ternary compound. Reprinted, by permission, from Phase Diagrams for Ceramists, Vol. 1, p. 15. Copyright  1964, The American Ceramic Society.

these systems are carbides, nitrides, and oxides, respectively. For example, consider the copper–iron–oxygen (Cu–Fe–O) phase diagram shown in Figure 2.20. We know that this must be an isothermal section of a more extended diagram, such as that in Figure 2.19, and in this case it turns out to be a range of temperatures at which all the phases indicated are stable, namely, above 675◦ C. Along the Fe–O binary, we find common ferrous oxides such as hematite (Fe2 O3 ), magnetite (Fe3 O4 ), and w¨ustite (FeO). The lines that connect these intermediate binary compounds with the third component are called alkemade lines, or alkemades. The alkemades divide ternary systems into composition triangles. For example, the alkemade connecting magnetite with Cu in Figure 2.20 could serve as the edge of a Fe–Cu–Fe3 O4 phase diagram. In practice, any compound can be used as the vertex of the triangle. Such diagrams that illustrate the phase relationships between important compounds are called subsystem diagrams.

174

THERMODYNAMICS OF CONDENSED PHASES

O

CuFe2O4 Tenarite

Delafossite (CuFeO2)

Hematite Magnetite Wüstite

Cuprite

Cu

Atom %

Fe



Figure 2.20 Stable assemblages above 675 C for the system Cu–Fe–O. Reprinted, by permission, from Phase Diagrams for Ceramists, p. 30, Fig. 2136, 1969 Supplement. Copyright  1969, The American Ceramic Society.

If we were to isolate the portion of the diagram along the Fe–O binary and expand the phase diagram along the temperature axis, we would end up with a binary component phase diagram like that shown in Figure 2.21, which begins at the compound FeO (the phases are mostly the same from pure Fe to the compound FeO), then continue up to Fe2 O3 , which corresponds to about 30 wt% O. This is, in fact, how many oxide ceramic phase diagrams are presented—that is, not with oxygen as a continuously varying independent element, but as an integral part of metal oxide compounds that form when it reacts with the metallic components–in this case, ferrous oxides. After all, pure oxygen is in the gaseous form at the temperatures under consideration here. As a result, phase diagrams such as that in Figure 2.21 are often presented at a total pressure of 1 atm, with oxygen partial pressures superimposed on it in the form of isobars, or lines of constant partial pressure. The isobars help us follow the variation in equilibrium compositions with temperature under a fixed oxygen partial pressure in the system. This is important from an industrial standpoint, where the partial pressure can be controlled in manufacturing operations. For example, consider a small quantity of liquid iron at 1635◦ C, saturated with oxygen in a gas reservoir of oxygen partial pressure, po2 = 10−8 atm. The volume of the gas reservoir is sufficiently large that any oxygen consumed by oxidation or produced by reduction of the iron or oxides has no effect on the total pressure of the system. The oxygen-saturated liquid iron, located at point a in Figure 2.21, is cooled slowly, such that equilibrium is maintained

175

THERMODYNAMICS OF CERAMICS AND GLASSES

a iron + wüstite

liquid 10 −6

10 −4

magnetite + liquid

wüstite + liquid

b

10−6

c

10−8

d

10−12 wüstite + magnetite

wüstite

10−14

10−16

f

10−18 10−20 10−22

600 10−28 a iron + magnetite

400

10

20

30

40

10−24 10−26 10−30

60 FeO·Fe2O3 80

50

hematite + liquid hematite 100 Air

10−2 magnetite + hematite

e

10−10

800

200 FeO

10 0

10 −2

A

magnetite

a 10 −8

T (°C)

liquid iron + liquid oxide 1600 a iron + liquid 1400 g iron + liquid 1200 g iron + wüstite 1000

10−8 10−10

g

10−12 10−14 10−16 10−18 10−20

90

Fe2O3

Weight %

Figure 2.21 The FeO–Fe2 O3 phase diagram showing oxygen isobars (in atm). Adapted from W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

with the gas reservoir. Since we are cooling at a constant oxygen partial pressure, we must follow along the isobar, represented by the dotted line a –b. We are still in the single-phase liquid region, but the dissolved oxygen is now beginning to react with the iron to form FeO: Fe(l) + 21 O2 (g)

FeO(l)

(2.45)

At point b, the liquid has mostly turned to w¨ustite, FeO, which is in the solid phase. From point b to point c, the remaining liquid becomes solid w¨ustite. A further decrease in temperature leads through the w¨ustite solid-phase field along the 10−8 atm isobar, until d is reached, at which point magnetite, Fe3 O4 , begins to form from the following oxidation reaction: 3FeO(s) + 12 O2 (g)

Fe3 O4 (s)

(2.46)

This reaction proceeds at 1275◦ C along the isobar to point e on the phase diagram. At this point, only magnetite remains. Magnetite remains a stable compound upon further cooling, to point f . Here, the formation of hematite, Fe2 O3 , becomes thermodynamically favorable at this partial pressure: 2Fe3 O4 (s) + 12 O2 (g)

3Fe2 O3 (s)

(2.47)

This reaction proceeds at 875◦ C until hematite is the only remaining component at point g.

176

THERMODYNAMICS OF CONDENSED PHASES

Equations (2.45) through (2.47) are typical oxidation reactions for the formation of metal oxides. The reverse reactions (steps g –a) are the reduction of metal oxides to form metal, such as is found in smelting operations. It should be apparent that the control of the oxygen partial pressure during heating is an important parameter in determining which phases will form. As before, the free energy of mixing of the various components in the system are calculated, and the equilibrium criteria applied to determine which phases are stable at the indicated temperature. An analysis similar to the one just performed for the iron–oxygen system can be performed for the copper–oxygen binary system in Figure 2.20, for which there are two stable copper oxides, cuprite (cuprous oxide, Cu2 O) and tenarite (cupric oxide, CuO). Further inspection of Figure 2.20 shows that copper oxide and iron oxide form two stable compounds: CuFe2 O4 and CuFeO2 . Recall that we can take vertical “slices” of the ternary phase diagrams, as well as isothermal (horizontal) slices. If we take, for example, a slice that begins at the tenarite composition (CuO) and extends across to the hematite composition (Fe2 O3 ), we would end up with a pseudobinary phase diagram, which, when plotted on the appropriate temperature–composition axes, would look like Figure 2.22. Note that the compound CuFe2 O4 is present, here labeled as spinel (see Section 1.2.2.3), but there is much more phase and temperature information available to us. This is, in fact, how many metal oxide phase diagrams are presented. The most stable forms of the

1600

Fe3O4

1458°

Sp

ine

Liquid

1400

Liq. + Spinel ss

1200

Spinel ss + Fe2O3

Liq. + CuO

1000

Fe2O3

T (°C)

ls

s

1575°

CuO + Spinel ss CuFe2O4

800

CuO

20

40

60

Mol. %

80

Fe2O3

Figure 2.22 Binary component phase diagram for the system CuOFe2 O3 . Reprinted, by permission, from Phase Diagrams for Ceramists, p. 31, Fig. 2140, 1969 Supplement. Copyright  1969 by The American Ceramic Society.

THERMODYNAMICS OF CERAMICS AND GLASSES

177

oxides, or at least the oxides of interest, are selected for the end compositions of the phase diagram, rather than plotting the elemental metals, along with oxygen, at the vertices of an equilateral triangle. In this way, a ternary phase diagram composed of three elements (Figure 2.20) can be represented as a binary phase diagram of two compounds (Figures 2.21 and 2.22). This type of representation allows us to do something even more powerful. We can now take what would normally be a quaternary component system, composed of four elements, and represent it on a ternary phase diagram system by representing three of the elements as important compounds of the fourth. Again, this is done primarily with metal oxide systems, in which three of the components are metals and the fourth is oxygen. One of the most important ternary component phase diagram systems represented in this way is for the system MgO–Al2 O3 –SiO2 , shown in Figure 2.23. This system is of great industrial importance because of the wide variety of naturally occurring compounds it contains. Silica, SiO2 , is found in nature in many forms, including sand and quartz. Alumina, Al2 O3 , is found as bauxite. Notice that we are using a new naming convention. Compounds that contain metallic elements as their root and end in “-a” are common industrial terms for oxides of that metal. Magnesia (MgO), yttria (Y2 O3 ), and zirconia (ZrO2 ) are a few more examples. This terminology is not

SS

Al2O3 1925 ± 40° 2050 ± 20°

MgO•Al2O3 (spinel) 2135 ± 20°

3Al2O3•2SiO2 (mullite)

Co

ru

4MgO•5Al2O3•2SiO2 (Sapphirine)

nd

um

1810 ± 10°

2030 ± 20° Spinel 1482 ± 3°

1575 ± 5° 1460 ± 5°

Sapphirine field

2MgO•2Al2O3•5SiO2 Kaolin (dehydrated) Cordierite ceramics

Mullite C 1370 ± 5° Cordierite °1443 ± 5° 5 5± e Periclase 1360 ± 5° 134 mit y ° id 10 Tr 1700 ± 25° ± 0 7 14 Cristobalite Forsterite B A 1470 ± 10° 1557 ± 2° 1543 ± 2° D 1850 ± 20° Two liquids

ensta

Proto

Magnesia refractories

tite

1453 ± 5°

MgO 2800°±

1595 ± 10°

1695 ± 5° 1695 ± 5° SiO2 2MgO•SiO2 MgO•SiO 2 Talc (dehydrated) 1713 ± 5° (forsterite) (enstatite) Steatites 1890 ± 40° Forsterite ceramics Low-loss steatites

Figure 2.23 The MgO–Al2 O3 –SiO2 phase diagram. Temperatures are in ◦ C. From Introduction to Ceramics. W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

178

THERMODYNAMICS OF CONDENSED PHASES

Example Problem 2.2

Ytrrium–barium–copper oxide (YBCO) superconductors were one of the first class of metal oxides to be found to superconduct (see Section 6.1.2.4) at liquid nitrogen temperatures (above about 77 K). Not all of the compounds of YBCO will superconduct, however, owing to the complex dependence of this phenomenon on the copper oxide lattice. For example, the “2–1–1” phase, named after the relative molar quantities of the three metals, Y2 BaCuO5 , does not superconduct, but the “1–2–3” phase, YBa2 Cu3 O7 , does. A portion of the Y2 O3 –BaO–CuO ternary phase diagram is shown below, with the corresponding pseudobinary phase diagram connecting the 1–2–3 compound with point F in the ternary phase diagram next to it. Determine the amount of 1–2–3 formed when a liquid of composition shown in the hatched region of the binary diagram (about 88 mol% BaCu3 O4 ) is cooled to the vicinity of 960◦ C. Answer: Application 1300 of the lever rule in the 1–2–3 + Liquid region Liquid shows that even in the 1200 BaO (BaCO ) best case (at the left edge of the hatched region) 1100 2:1:1 + Liq. only about 3% of the Ba Y O Ba CuO 1–2–3 phase is formed, BC the rest being liquid Ba Y O BaCuO + 1000 2:1:1 Ba Y O Liq. 2:1:1 + 1:2:3 + Liq. of different composition. Vicinity of BaY O 1:2:3 + Liq. 960 °C 1:2:3 As the liquid is cooled 1:2:3 + BC + Liq. Y BaCuO F YBa Cu O 900 BC + CuO+ Liq. further, more yttria-poor CuO Y O 1:2:3 + BC + CuO compounds are formed, 60 80 100 0 20 40 0 20 40 60 80 100 Y Cu O ) CuO Mol% 1/6(Ba YCu O ) 1/4(BaCu O ) such as barium cuprate 1/2(Y O ) Mol% (BC), that are difficult to remove. Thus, melt processing of pure 1–2–3 superconductors is virtually impossible via this route. One solution has been to use nonequilibrium processing, which renders the equilibrium phase diagrams shown above useless. Reprinted, by permission from T. P. Sheahen, Introduction to High Temperature Superconductivity, pp. 179, 180. Copyright  1994 by Plenum Publishing. T, °C

3

4 2 7

2

2 2 5

3

2

3 4 9

2 4

2

5

2

3 7

2 3

2 3

2

2 5

2

3 7

3 4

universal, especially when the metal can have more than one common oxidation state. Many of the compounds along the Al2 O3 –SiO2 binary, termed aluminosilicates, are the building blocks of most clays. Kaolin, for example, is a hydrated aluminosilicate of nominal composition Al2 O3 ·2SiO2 ·2H2 O, sometimes expressed as the compound Al2 Si2 O5 (OH)4 . It is used for making porcelain for valves, tubes, and fittings, as a refractory for bricks and furnace linings, and as a pigment filler in paints. Talc is a hydrated form of magnesium silicate, 3MgO·4SiO2 ·H2 O, and can be dehydrated to form protoenstatite crystals, MgSiO3 , in a silica matrix. Cordierite ceramics have variable compositions as shown in the shaded region of Figure 2.23, but have a nominal composition of 2MgO·2Al2 O3 ·5SiO2 . Cordierite ceramics are particularly useful due to their low coefficient of thermal expansion (see Section 5.1.3.1) and thermal shock resistance. Many of the ceramics in this system are termed refractories, since they have very high melting points, generally above 1580◦ C, and are able to bear loads and resist corrosion at elevated temperatures.

THERMODYNAMICS OF CERAMICS AND GLASSES

179

2.2.1.3 Ellingham Diagrams∗ . Equation (2.45) is an example of a formation reaction involving solid and gas reactants, leading to the formation of a solid metal oxide. The explicit free energy relations for this reaction were eliminated from the previous discussion for two principal reasons: (1) because the connection between thermodynamics and reaction kinetics is dealt with in Chapter 3 and (2) because the reaction involves a gaseous species, and we have purposely limited our discussion to condensed phases to this point. However, the formation of metal oxides by oxidation of their metallic precursors and its associated thermodynamics are important topics in both ceramics and physical metallurgy, such that a brief elaboration is in order. To do so effectively, we will need to discuss the thermodynamics of gases and introduce a quantity called the thermodynamic activity, or just activity. You may wish to look ahead to Chapter 3 for further discussion on reaction kinetics. The goal of this section is to develop a straightforward graphical representation of oxidation reactions that will allow us to determine which oxides form most favorably in mixtures of metals. Consider the reaction equilibrium between a pure solid metal, M, its pure oxide, MO, and oxygen gas at temperature T and pressure P :

2M(s) + O2 (g)

2MO(s)

(2.48)

It can be shown (see Chapter 3) that the standard state Gibbs free energy change for this reaction is given by the difference between the standard state free energies of the products minus the reactants and can be related to the partial pressure of the gaseous species: ◦







G = 2GMO(s) − GO2 (g) − 2GM(s) = −RT ln PO−1 2

(2.49)

where PO2 is the partial pressure of oxygen and R is the gas constant. Thus, in the case of reaction equilibrium involving only pure condensed phases and a gas phase, the Gibbs free energy change can be written solely in terms of the species that appear only in the gas phase. (You may recall from thermodynamics that this is due to that fact that the thermodynamic activity of condensed phases is usually taken as unity.) At any fixed temperature, the establishment of reaction equilibrium occurs at a unique value of oxygen partial pressure, which we will designate PO2 ,eqT . The equilibrium thus has one degree of freedom, as given by the phase rule (φ = 3, two pure solids and a gas; C = 2, metal + oxygen; N = 2, temperature and pressure), F = C − φ + N = 2 − 3 + 2 = 1. At any temperature T , when the actual partial pressure of oxygen in a closed metal–metal oxide–oxygen system is greater than PO2 ,eqT , spontaneous oxidation of the metal will occur, thereby consuming oxygen and decreasing the oxygen partial pressure in the gas phase. When the actual oxygen pressure has been decreased to PO2 ,eqT , the oxidation ceases and reaction equilibrium is reestablished. Similarly, if the oxygen pressure in the closed vessel was originally less than PO2 ,eqT , spontaneous reduction of the oxide would occur [the reverse of Eq. (2.48)] until PO2 ,eqT was reached. Metallurgical processes involving the reduction of oxide ores rely on the establishment and maintenance of an oxygen pressure less than PO2 ,eqT in the reaction vessel. The variation of the standard state Gibbs free energy change for the oxidation reaction at any temperature from experimentally measured variations in PO2 ,eqT can be fitted to an equation of the form: ◦

G = A + BT ln T + CT

(2.50)

180

THERMODYNAMICS OF CONDENSED PHASES

For example, let us consider the oxidation of Cu to Cu2 O: 4Cu(s) + O2 (g)

2Cu2 O(s)

(2.51)

The empirically fitted free-energy–temperature relationship (in joules) is ◦

G (J ) = −333,900 + 14.2 ln T + 247T

(2.52)

which can be approximated in linear form as ◦

G (J ) = −333,000 + 141.3T

(2.53)

in the temperature range 298 to 1200 K. The error in using Eq. (2.53) instead of Eq. (2.52) is approximately 0.3% at 300 K and 1.4% at 1200 K. The linear approximation allows us to make a series of simple constructions that compare the free energy of formations for a variety of oxidation reactions. H. J. T. Ellingham plotted the experimentally determined variations of G◦ with temperature for the oxidation of a series of metals, and he found that the linear approximation was suitable when no change of state occurred. Thus, all free energy expressions for the oxidation of metals could be expressed by means of a simple equation of the form ◦

G = A + BT

(2.54)

where the constant A is identified with the temperature-independent standard enthalpy change for the reaction, H ◦ , and the constant B is identified with the negative of the temperature-independent standard entropy change for the reaction, S ◦ . Consider the variation of G◦ with T for the oxidation of silver: 4Ag(s) + O2 (g)

2Ag2 O(s)

(2.55)

The plot of free energy with temperature for this reaction, using the correlation provided by Eq. (2.54), is called an Ellingham diagram and is shown in Figure 2.24. At 462 K, G◦ = 0, and pure solid silver and oxygen gas at 1 atm pressure are in equilibrium with silver oxide at this temperature. If the temperature is decreased to T1 , G◦ < 0, the metal phase becomes unstable relative to silver oxide and oxidizes spontaneously. The value of PO2 ,eqT at this temperature is calculated using Eq. (2.49), and since G◦ is a negative quantity at this temperature, PO2 ,eqT < 1 atm. Similarly, if the temperature of the system is increased to T2 , then G◦ > 0, and the oxide phase becomes unstable relative to silver metal and oxygen gas at 1 atm. In this instance, PO2 ,eqT > 1 atm. The value of G◦ is thus a measure of the chemical affinity of the metal for oxygen, and the more negative its value at a given temperature, the more stable the oxide. Ellingham diagrams can be made for any metal–metal oxide combination. When plotted together, the graphical representation of free energy versus temperature provides a useful method for evaluating the relative affinity of oxygen for metals. By way of example, let us first consider two general oxidation reactions for which the Ellingham lines intersect each other, as shown in Figure 2.25: 2A(s) + O2 (g) B(s) + O2 (g)

2AO(s) BO2 (s)

(2.56)

THERMODYNAMICS OF CERAMICS AND GLASSES

181

→ g) 2(

O (s )

+

10 4A g

∆G °, kJ/mol of oxygen

T2

2A g

T1

20

2O (s )

30

0

−10

462 K

−20 300

400

500 T (K)

600

700

∆G °

Figure 2.24 Ellingham diagram for the oxidation of silver. Reprinted, by permission, from D. R. Gaskell, Introduction to Metallurgical Thermodynamics, 2nd ed., p. 272. Copyright  1981 by McGraw-Hill Book Co.

B+

O2



BO 2

2A + O 2

0K

→ 2AO

TE T (K)

Figure 2.25 Ellingham diagram for two hypothetical oxidation reactions. Reprinted, by permission, from D. R. Gaskell, Introduction to the Thermodynamics of Materials, 3rd ed., p. 360. Copyright  1973 by Taylor & Francis.

The two Ellingham plots intersect at temperature TE . At temperatures less than TE , A and BO2 are stable with respect to B and AO. At temperatures above TE , B and AO are stable. At TE , A, B, AO, and BO2 are in equilibrium with each other. Thus, if pure A were to be used as a reducing agent to reduce pure BO2 to form B and pure AO, then the reduction would have to be conducted at temperature higher than TE . This is the

182

THERMODYNAMICS OF CONDENSED PHASES

power of the Ellingham diagram: Not only can it be used to determine the stability of a single oxide at a given temperature, but it can provide relative stabilities of various oxides and indicate which metals can be used as reducing agents for a given metal oxide. You should note that in order to compare the stabilities of different oxides, the Ellingham diagrams must be drawn for oxidation reactions involving the consumption of the same number of moles of oxygen. The units of G◦ for the oxidation reaction must then be given in energy per mole of O2 consumed. In order to avoid calculating the value of PO2 ,eqT for all oxidation reactions, F. D. Richardson added a nomographic scale along the right-hand edge of the Ellingham diagram. The value of PO2 ,eqT for any metal–metal oxide equilibrium is read off the graph as that value on the scale which is collinear with the points G◦ = 0, T = 0, and G◦T , T = T . Ellingham diagrams with this additional nomograph are often called Ellingham–Richardson diagrams. An Ellingham–Richardson diagram for the formation of common metal oxides is shown in Figure 2.26. Finally, it should be noted that phase changes can be accommodated in the Ellingham diagram. When the temperature moves above the melting point of the metal or metal oxide, their corresponding standard states must change. For example, above 1100◦ C, copper metal is no longer solid, and the oxidation reaction of interest is: 4Cu(l) + O2 (g)

2Cu2 O(s)

(2.57)

The Ellingham diagram reflects this change in state with a change in slope, albeit a small one for this example, above the melting point of copper, indicated by M. Melting points of all the metals are so indicated, and the melting points of the respective oxides with a boxed M. Boiling points of the metals, where appropriate, are indicated with a B, and the free energy line once again may undergo a change in slope. No metal oxides reach their boiling points at the temperatures of interest in the diagram. The free energy line for the formation of carbon monoxide and carbon dioxide are included in this plot, as the reduction of metal oxides with carbon monoxide is common in smelting operations: MO2 + 2CO

M + 2CO2

(2.58)

Similar Ellingham–Richardson diagrams for the formation of nitrides and sulfides can be constructed and are available in the literature. 2.2.2

Interfacial Thermodynamics

Thermodynamics plays a fundamental role in practically all of the processes that are described in this book. Phase equilibria, which has been the focus of this chapter so far, is but one of these areas. Let us diverge from phase equilibria for a bit, and discuss a thermodynamic topic that will be of use in many of the subsequent chapters: interfacial energy. Interfacial energies will be developed using thermodynamic arguments primarily for liquids, but we will see that the results are general to solids as well. This point will be illustrated by application of interfacial energies to an important process in the densification of commercial materials called sintering. Later on, we will see that it has enormous utility in the description of many surface-related phenomena in materials science, such as fracture, adhesion, lubrication, and reaction kinetics.

183

THERMODYNAMICS OF CERAMICS AND GLASSES

pH / pH O 2 2

0

−8

−7

10

pCO/pCO 2

−14

10 2Ag 2O = O + 2 4Ag

−12

10

−10

10

−9

−8

10 10

−7

10

−6

10

10

−6

−5

10 −5

10

−1

10

M

O 2Cu 2

4C

−300 −400

∆G ° = RT In p O2 (kJ)

e+

3/2F

−500

O2

C −600 r + O2

M

= + O2 6Fe M 2CO 2 2NiO o + O 2 = O2 = O2 = C 2H 2O 2 + + = i N O 2 + O2 2C 2H 2 M 2FeO M O2 = + C e + O = CO F 2 2 2

2C

e 3O 4

M

+O

2Mn

O3 3Cr 2 = 2/

=

1

−6

10 −8

10

10 2

10 −10

10

2

−12

M

O 2Mn

M

M

10

6

10

10

=

= TiO

2

7

−20

10

−22

10

−24

109

10

l + O2

Al 2O 3

3

= 2/

B

M

B

10

4/3A

−1000

M

gO

g+

−1100

2M

O2

M =2

+ 2Ca

O2

M

=2

M

Melting point of metal

B

Boiling point of metal

M

Melting point of oxide

200

pO , (atm) 2 pCO/pCO 2 pH / pH O 2 2

400

600 −100

10

−80

10

10

1000 10

10

−42

10

−38

10

8

10

9

10 −28

10

11

10

10

10 12

10

1200 1400 1600 −50

7

10

−30

−60

6

10

10

10

800

5

10

10

−26

CaO

temperature, °C

0

4

8

10

−900

0K

5

10 −16

−18

−800

−1200

10

10

SiO 2

O2 Ti +

3

4

10

−14

10

M

4/3C

O2

10

10

−700 Si +

−1

10

3

O

+ O2

10 1

10

= 2C

2

10−2

−3

2CoO

3F = 1/

−1

10

10 O4 −4 10 2Fe 3

M

H

−2

10

1 10−2

= u + O2

10

10 −3

10

10

−100 −200

−3

−4

10 −4

11

−34

10

13

10

10

1014 10

13

1012

Figure 2.26 Ellingham–Richardson diagram for some common metal oxides. Reprinted, by permission, from D. R. Gaskell, Introduction to the Thermodynamics of Materials, 3rd ed., p. 370. Copyright  1973 by Taylor & Francis.

2.2.2.1 Surface Energy. A surface is an inhomogeneous boundary region between two adjacent phases. As shown in Figure 2.27, atoms on the surface of a phase are necessarily different than those in the bulk. In particular, they have fewer nearest neighbors than the bulk, and they may be exposed to constituents from an adjacent phase. This generally means that less energy is required to remove an atom from a surface than to remove it from the bulk. Therefore, the potential energy of surface atoms is higher than bulk atoms. In turn, work is required to move atoms from the bulk to the surface. When this is done, new surface is created, and the surface area of the phase increases. The reversible work required to form the new surface, dW s , is

184

THERMODYNAMICS OF CONDENSED PHASES

Figure 2.27 Schematic representation of surface and bulk atoms in a condensed phase. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

proportional to the surface area, dA, that is created: dWS = γ dA

(2.59)

The proportionality constant, γ , is called the surface energy. When the bulk phase in question is a liquid, the surface energy is often called surface tension. Rearranging Eq. (2.59) gives: dWS γ = (2.60) dA We can see from this relation that surface energy is work per unit area, so it should have units of J/m2 in SI and ergs/cm2 in cgs. Often you will see values of surface energy expressed in units of dyne/cm, where 1 dyne = 1 erg/cm. With the aid of Eqs. (2.1), (2.7), (2.9), and (2.11), you should be able to prove to yourself that the reversible, non-pressure–volume work, dW s , is equivalent to the free energy change, dG, so that Eq. (2.60) becomes, with proper use of partial differentials,   ∂G (2.61) γ = ∂A T ,P ,Ni This relationship identifies the surface energy as the increment of the Gibbs free energy per unit change in area at constant temperature, pressure, and number of moles. The path-dependent variable dW s in Eq. (2.60) has been replaced by a state variable, namely, the Gibbs free energy. The energy interpretation of γ has been carried to the point where it has been identified with a specific thermodynamic function. As a result, many of the relationships that apply to G also apply to γ : γ = Hs − T Ss

(2.62)

THERMODYNAMICS OF CERAMICS AND GLASSES

185

where the subscript s on the enthalpy and entropy indicates that these are surface properties. As mentioned earlier, the surface atoms are fundamentally different than bulk atoms, so that they have different enthalpies and entropies associated with them. Differentiation of Eq. (2.62) with respect to temperature at constant pressure gives 

∂γ ∂T



P

= −Ss

(2.63)

Substitution of Eq. (2.63) in (2.62) gives γ = Hs + T



∂γ ∂T



(2.64)

P

Equation (2.64) is useful from an experimental standpoint because the measurements of surface energies at various temperatures can, in principal, provide a measurement of the surface enthalpy. The surface enthalpy, Hs , can also be determined directly, because it is equivalent to the heat of sublimation or vaporization. Since the surface energy is a direct result of intermolecular forces, its value will depend on the type of bond and the structural arrangement of the atoms. Normally, densely packed planes would have lower surface energies. Liquid hydrocarbons having only weak van der Waals forces have values of surface tension in the neighborhood of 15–30 dyn/cm, while liquids with polar forces and hydrogen bonds, like water, have surface tensions in the range of 3–72 dyn/cm. Fused salts and glasses with additional ionic bonds have surface energies from 100 to 600 dyn/cm, and molten metals have surface tensions from 100 to 3000 dyn/cm. Even higher surface energies can be found in covalent solids (see Example Problem 2.3). See Appendix 4 for values of solid and liquid surface energies for a variety of materials. 2.2.2.2 The LaPlace Equation. The concept of surface energy allows us to describe a number of naturally occurring phenomena involving liquids and solids. One such situation that plays an important role in the processing and application of both liquids and solids is the pressure difference that arises due to a curved surface, such as a bubble or spherical particle. For the most part, we have ignored pressure effects, but for the isolated surfaces under consideration here, we must take pressure into account. Consider a generic curved surface such as that found in a sphere or cylinder (see Figure 2.28). The curved surface has two principal radii of curvature, R1 and R2 . The front of this surface is indicated by the line xx 1 , and the back is represented by the line yy1 . Let us now move the surface out by a differential element, dz. The front of the new surface is now given by x˙ x˙1 , and the back is indicated by y˙ y˙1 . The work required to displace the surface this amount is supplied by a pressure difference, P . The pressure acts on an area given by (xx 1 )(x1 y1 ) moving through the differential element dz, such that the total pressure–volume work associated with extending the surface through dz is W = P (xx1 )(x1 y1 ) dz (2.65)

The pressure–volume work must be counterbalanced by surface tension forces. The work required to move against surface tension forces is best calculated by breaking it

186

THERMODYNAMICS OF CONDENSED PHASES

Example Problem 2.3

An alternative method for estimating surface energies is to calculate the work required to separate two surfaces of a crystal along a certain crystallographic plane. At 0 K, this work can be approximated as the energy required to break the number of bonds per unit area, or the energy of cohesion, Ecoh . Since two surfaces are being formed in this cleavage process, the surface energy of a single surface is then γ ≈ 1/2Ecoh Let us consider, for example, separation along the (111) plane in diamond. The lattice ˚ (see Table 1.11), so the number of atoms per square constant for diamond is a = 3.56 A centimeter on the (111) surface is 4 2 atoms per plane(111) √ = √ 2 3a /2 3(3.56 × 10−8 )2

= 1.82 × 1015 atoms/cm2

From Appendix 1, the bond energy for C–C bonds is 348 kJ/mol (83.1 kcal/mol), so that the cohesive energy is Ecoh =

(1.82 × 1015 bonds/cm2 )(348,000 J/mol)(107 erg/J) 6.02 × 1023 atoms/mol

= 10,500 erg/cm2 so that the surface energy is

γ = 1/2(10,500 erg/cm2 ) = 5250 erg/cm2

into two parts, W1 and W2 . W1 is the work required to move side xx1 away from yy1 a distance (x1 y1 /R1 ) dz during the expansion: W1 =

γ (xx1 )(x1 y1 ) dz R1

(2.66)

Similarly, the work required to move side xy away from x1 y1 through a distance (xx 1 /R2 ) dz is W2 : W2 =

γ (x1 y1 )(xx1 ) dz R2

(2.67)

Adding Eqs. (2.66) and (2.67) together to arrive at the work against surface tension, equating them with the pressure–volume work in Eq. (2.65) and simplifying leads to the Laplace equation:   1 1 P = γ + (2.68) R1 R2

THERMODYNAMICS OF CERAMICS AND GLASSES

y· y x· dz

x1

x

y· 1 y1

R1

x· 1

187

O′

O

R2

Figure 2.28 An element of a curved surface with principal radii R1 and R2 . Reprinted, by permission, from J. F. Padday, in Surface and Colloid Science, E. Matijevic, ed., Vol. 1, p. 79. Copyright  1969 by John Wiley & Sons, Inc.

The Laplace equation in this form is general and applies equally well to geometrical bodies whose radii of curvature are constant over the entire surface to more intricate shapes for which the Rs, are a function of surface position. In the instance of constant radii of curvature across the surface, Eq. (2.68) reduces for several common cases. For spherical surfaces, R1 = R2 = R, where R is the radius of the sphere, and Eq. (2.68) becomes: 2γ (2.69) P = R For a cylindrical surface, R1 (or R2 ) is infinity, so the remaining radius, R, is the radius of the cylinder and: P =

γ R

(2.70)

Finally, for a planar surface, both radii of curvature become infinity, and: P = 0

(2.71)

The pressure difference may also be numerically zero in the instance where the two principal radii of curvature lie on opposite sides of the surface, such as in the case of a saddle. 2.2.2.3 The Young Equation. The principle of balancing forces used in the derivation of the Laplace equation can also be used to derive another important equation in surface thermodynamics, the Young equation. Consider a liquid droplet in equilibrium

188

THERMODYNAMICS OF CONDENSED PHASES

gL Liquid g LS

q

gS

Solid

Figure 2.29 Schematic diagram of liquid droplet on solid surface. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed., Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

with its own vapor and a flat, solid surface at constant temperature, as shown in Figure 2.29. The liquid–solid, liquid–vapor, and solid–vapor interfacial surface energies are defined as γLS , γLV , and γSV , respectively. Technically, the liquid–vapor and solid–vapor interfacial energies should be for the liquid and solid in equilibrium with their respective vapors, which are probably not the same. In practice, however, the vapor is usually a nonreactive gas, so γLV and γSV become the surface tension and solid surface energy in the gas of interest and are simply labeled γL and γS , respectively, as shown in Figure 2.29. These are both approximations. The surface tensions and surface energies of some common liquids and solids are listed in Appendix 4. At equilibrium an angle θ , called the contact angle, is formed at the three-phase solid–liquid–gas junction. The contact angle can have values from zero to 180◦ . When θ = 0, the liquid completely spreads of the solid surface, forming a thin monolayer (see Figure 2.30). Such a condition is termed wetting. The other extreme is called nonwetting, and it occurs when the entire liquid droplet sits as a sphere on the solid. All values of contact angle in between these two extremes are theoretically possible. Obviously, the chemical nature of the liquid, solid, and vapor determines the extent to which the liquid will wet the solid, and temperature has an important influence. Keep in mind that the contact angle under consideration here is an equilibrium contact angle. There are also dynamic, receding, and advancing contact angles associated with droplets as they spread and move on substrates. For now, the term contact angle will refer to the equilibrium condition. The three interfacial surface energies, as shown at the three-phase junction in Figure 2.29, can be used to perform a simple force balance. The liquid–solid interfacial energy plus the component of the liquid–vapor interfacial energy that lies in the same direction must exactly balance the solid–vapor interfacial energy at equilibrium: γL cos θ + γSL = γS

(2.72)

This equation is called Young’s equation; it is named after Thomas Young, who first proposed it in 1805. The derivation presented here in terms of force balances is simplistic, but there are more rigorous thermodynamic arguments to support its development. In practice, the contact angle can be experimentally determined in a rather routine manner, as can the liquid surface tension and even the solid surface energy. The interfacial energy for the liquid–solid system of interest, γSL , can then be calculated using Young’s equation. Alternatively, if γSL , γL , and γS are known as a function of temperature, the contact angle can be predicted at a specified temperature.

THERMODYNAMICS OF CERAMICS AND GLASSES

189

Liquid 180° q

( a) Liquid

Solid

(b)

Figure 2.30 Schematic illustration of (a) nonwetting and (b) wetting of a liquid on a solid. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed., Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

2.2.2.4 Sintering and Densification. Curved surfaces and surface energies play a large role in the description of an important phenomenon in ceramic processing called sintering. Sintering is a process in which a particulate material is consolidated during heat treatment. Consolidation implies that within the material, particles have joined together into an aggregate that has strength. Densification, or the increase in density through the accompanying reduction in interparticle voids, often accompanies sintering, but there are some highly porous insulation products that can actually be less dense after they have been sintered. Nonetheless, we will examine sintering and densification simultaneously. We will also limit the discussion to solid state sintering. The driving force for sintering is the reduction in the total free energy of the particulate system, G, which is composed of free energy changes of volume, GV , boundaries, Gb , and surfaces, Gs :

G = GV + Gb + GS

(2.73)

In conventional sintering, the dominant term in the free energy reduction is that due to surface area reduction. Recall from Eq. (2.61) that the surface energy is the increment of the Gibbs free energy per unit change in area at constant temperature, pressure, and total number of moles   ∂G (2.61) γ = ∂A

190

THERMODYNAMICS OF CONDENSED PHASES

such that if As is the surface area of the particles, then the free energy change due to surface area is GS = γ As (2.74) Three stages of solid-state sintering are recognized, although there is not a clear distinction between them, and they overlap to some extent. ž

ž

ž

Initial sintering involves rearrangement of the powder particles and formation of a strong bond or neck at the contact points between particles (see Figure 2.31). The relative density of the powder may increase from 0.5 to 0.6 due mostly to an increase in particle packing. Intermediate sintering occurs as the necks grow, the number of pores decreases substantially, and shrinkage of the particle assembly occurs. Grain boundaries form, and some grains grow while others shrink. This stage continues while the pore channels are connected, called open porosity, but is considered over when the pores are isolated or in closed porosity. The majority of shrinkage occurs during intermediate sintering, and the relative density at the end of this stage is about 0.9. Final-stage sintering occurs as the pores become closed and are slowly eliminated by diffusion of vacancies from the pores along the grain boundaries. Little densification occurs in this stage, although grain sizes continue to grow.

(a)

( b)

(c)

(d )

Figure 2.31 Development of ceramic microstructure during sintering: (a) Loose powder particles; (b) initial stage; (c) intermediate stage; and (d) final stage. From W. E. Lee and W. M. Rainforth, Ceramic Microstructures, p. 37. Copyright  1994 by William E. Lee and W. Mark Rainforth, with kind permission of Kluwer Academic Publishers.

THERMODYNAMICS OF POLYMERS

Pore gs

191

gg

f Solid grain

Solid grain ggb

Figure 2.32

Schematic illustration of dihedral angle for solid–pore interaction.

During the initial stage, pore rounding causes a decrease in Gs in proportion to the reduction of the surface area, as dictated by Eq. (2.74), but the formation of a grain boundary causes an increase in Gb . The angle of intersection of the pore at the pore–grain boundary juncture (see Figure 2.32) is called the dihedral angle, φ, and is related to the relative interfacial energies of the grain boundary, γgb and the solid surface, γs :   γgb φ (2.75) = cos 2 2γs Values of φ range from about 105◦ to 113◦ in ceramics, implying that γgb /γs is 1.1 to 1.2, whereas in metals this ratio ranges from 0.25 to 0.5. There is much more to be said concerning sintering, especially with regard to the role that diffusion plays. Many of the models that describe the sintering processes are based on the diffusivity of solid species, which are addressed in Section 4.3. The practical aspects of the sintering of metals and ceramics are described in Sections 7.1 and 7.2, respectively. 2.3

THERMODYNAMICS OF POLYMERS

Polymers don’t behave like the atoms or compounds that have been described in the previous sections. We saw in Chapter 1 that their crystalline structure is different from that of metals and ceramics, and we know that they can, in many cases, form amorphous structures just as easily as they crystallize. In addition, unlike metals and ceramics, whose thermodynamics can be adequately described in most cases with theories of mixing and compound formation, the thermodynamics of polymers involves solution thermodynamics—that is, the behavior of the polymer molecules in a liquid solvent. These factors contribute to a thermodynamic approach to describing polymer systems that is necessarily different from that for simple mixtures of metals and compounds. Rest assured that free energy will play an important role in these discussions, just as it has in previous sections, but we are now dealing with highly inhomogeneous systems that will require some new parameters. 2.3.1

Solution Thermodynamics and Phase Separation

Most polymers are synthesized in solution. The solvent could be water, carbon tetrachloride, benzene, or any number of aqueous or organic solvents. Many polymers are

192

THERMODYNAMICS OF CONDENSED PHASES

also processed in solution. They can be cast as solutions, and the solvent can be allowed to evaporate to leave only the polymer. So, it is critically important to understand how a polymer molecule interacts with the solvent. Again, intermolecular forces will play an important role here, but, as always, free energy will be used to determine the stability of any system we can create at any temperature. Recall that the free energy of mixing for a binary component system is Gmix = Hmix − T Smix

(2.29)

and that for regular solutions, this relation becomes Gmix = αXA XB + RT (XA ln XA + XB ln XB )

(2.33)

The first term in Eq. (2.33) is the enthalpy of mixing, and the second term is the entropy of mixing. Let us examine each of these terms independently, beginning with the entropy of mixing. For solutions of a polymer in a solvent, it can be shown, again through statistical thermodynamic arguments, that the entropy of mixing is given by Smix = −R(XA ln vA + XB ln vB ) = −kB (NA ln vA + NB ln vB )

(2.76)

Here kB is Boltzmann’s constant, or the gas constant per molecule, R/N0 , where N0 is Avogadro’s number (or NA + NB for one mole of solution) and vA and vB are the volume fractions of solvent and polymer. Comparison of Eq. (2.76) with the entropy of mixing presented earlier in the chapter by Eq. (2.30) shows that they are similar in form, except that now the volume fractions of the components, vA and vB , are found to be the most convenient way of expressing the entropy change for polymers, rather than the mole fraction used for most small molecules. This change arises from the differences in size between the large polymer molecules and the small solvent molecules which would normally mean mole fractions close to unity for the solvent, especially when dilute solutions are being studied. It can also be shown with related statistical arguments that the enthalpy of mixing for a regular polymer–solvent solution is given by Hmix = kB T χNA vB = RT χvA vB

(2.77)

where NA is the number of solvent molecules, and vA and vB are the volume fraction of solvent and polymer molecules, respectively. Compare the second form of Eq. (2.77) with Eq. (2.31) and you will see that they have the same general form, except that the mole fractions have been replaced with volume fractions, just as for the entropy of mixing, and the interaction energy, α, is composed of the gas constant, R, temperature, T , and a new parameter, χ, which is called the Flory–Huggins interaction parameter, or just interaction parameter (do not confuse the interaction parameter with electronegativity, which goes by the same symbol!). Note that the interaction parameter is the interaction energy per solvent molecule per kB T , and is inversely proportional to temperature. It is zero for ideal mixtures (zero enthalpy of mixing), positive for endothermic mixing, and negative for exothermic mixing. The interaction parameter is an important feature of polymer solution theory. It is an indication of how well the polymer interacts with the solvent. “Poor” solvents have values of χ close to 0.5,

THERMODYNAMICS OF POLYMERS

193

Cooperative Learning Exercise 2.5 It can be shown that the volume fraction of polymer in solution, vB , can be related to the weight fraction of polymer in solution, XB , as follows:

vB =

XB (1/xn ) + XB [1 − (1/xn )]

where n is the number of repeat units in a polymer chain, or the degree of polymerization (cf. Section 1.3.5.1). (This is a good thing to sit down and prove to yourself sometime. Just recall that XB = NB /(NA + NB ) and vB /vA = xn NB /NA .) For simplicity, let us assume a monodispersed polymer of xn = 500; that is, all the polymer chains have exactly 500 repeat units. First, work together to create a table of weight fractions in increments of 0.1 from 0.1 to 0.9; for example, XB = 0.1, 0.2, and so on, and their corresponding volume fractions using the equation above. Then perform the following calculations separately. Person 1: Use Eq. (2.30) to calculate values of Smix /R for the nine values of XB and XA . These values are the entropy of mixing for an ideal solution. Person 2: Use Equation (2.76) to calculate values of Smix /R for the nine values of vB and vA . [Hint: Use the first form of Equation (2.76).] These values are the free energy of mixing for a real “athermal” polymer solution. Combine your information to make a single, rough plot of Smix /R versus XB . At what weight fraction is the difference in the entropy of mixing between an ideal and athermal polymer solution the greatest for this system? Adapted from P. Hiemenz, Polymer Chemistry , p. 520. Copyright  1984 by Marcel Deker, Inc. Xn 0

0.2

0.4

0.6

0.8

1.0

0.5 1.0

ideal

1.5

R ∆Smix

2.0 2.5 3.0 3.5

n = 500

4.0

Answer : XB = 0.1; XA = 0.9; vB = 0.982; vA = 0.018; max difference occurs about XB = 0.2. Polymer solutions are highly nonideal.

while “good” solvents have lower, or negative, values. The typical range for most synthetic polymer solutions is 0.3 < χ < 0.6 (see Table 2.5). A linear temperature dependence for χ is also predicted of the form χ = a + b/T , which suggests that as the temperature increases, the solvating power of the liquid should increase.

194

THERMODYNAMICS OF CONDENSED PHASES

Table 2.5

Some Polymer–Solvent Interaction Parameters at 25◦ C

Solvent

Interaction Parameter χ

Toluene (V1 = 106)a Benzene (V1 = 89.0) Toluene Cyclohexane (V1 = 108) Benzene Cyclohexane

0.391 0.437 0.557 0.436 0.442 0.489

Benzene Benzene Benzene

0.390 0.486 0.564

Polymer cis-Polyisoprene Polyisobutylene Butadiene-styrene 71.5:28.5 Butadiene-acrylonitrile 82:18 70:30 61:39 a

Note that V1 is in cubic centimeters per mole. Source: F. Rodriguez, Principles of Polymer Systems, 2nd ed. Copyright  1982 by McGraw-Hill Book Company.

Combining now the enthalpy of mixing and the entropy of mixing, along with substituting them into Eq. (2.29), gives us the free energy of mixing for a polymer solution: Gmix = kB T [χNA vB + NA ln vA + NB ln vB ] (2.78) From this expression, one can derive many useful relations involving experimentally obtainable quantities, such as osmotic pressure. We will use it to develop a phase diagram for a polymer solution. Consider a binary component system consisting of a liquid, A, which is a poor solvent for a polymer, B. We already know that complete miscibility occurs when the Gibbs free energy of mixing is less than the Gibbs free energies of the components, and the solution maintains its homogeneity only as long as Gmix remains less than the Gibbs free energy of any two possible coexisting phases. This situation is shown in Figure 2.33. As with our other phase diagrams, the free energy is plotted as a function of composition in the top part of the curve, and the resulting temperature–composition plot is shown in the bottom of the diagram. At high temperatures, T5 , the polymer and solvent are miscible, and a single phase is formed (Region I). As T decreases, the solution begins to separate into two phases. Phase separation, or immiscibility, begins to occur at the critical temperature, Tc . Here, the free energy curve begins to have maxima, minima, and inflection points, and the condition of phase stability as given by Eq. (2.36), namely, that the chemical potentials of the two phases (here labeled L1 and L2) be equal, is no longer met:     ∂Gmix ∂Gmix µL1 = = µL2 = (2.36) ∂vB T ,P ,nL2 ∂vB T ,P ,nL1 (You should prove to yourself that it makes no difference if we differentiate with respect to component A or B, nor whether we differentiate with respect to mole or volume fraction, since the derivatives are set equal to zero.) The locus of points swept out by the stability condition shown in Eq. (2.36) at various temperatures is called the binodal, or cloud point curve. For example, just as we drew tangents to the

THERMODYNAMICS OF POLYMERS

∂2G =0 ∂v 2 T.P.

195

T5

free energy

Tc T3 T2 T1

vBI

temperature

stable region

spinodal locus

vBII

vB

critical point

I

II

binodal locus

vBIII

unstable region III

metastable region

T1

Concentration

Figure 2.33 Polymer–solvent phase diagram showing binodal, spinodal, and miscibility gap. Reprinted, by permission, from J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed., P. 167. Copyright  1991 by Chapman & Hall.

minima of the free energy curves in Figures 2.3 and 2.6 to map out the phase boundaries in solid–liquid equilibria, we can use tangents to the minima (line A–B) in a single free energy–composition plot for our polymer solution to establish the phase boundary—in this case, the binodal. A similar mapping of the inflection points, mathematically described by the second derivative of the free energy versus composition curve given in  2  ∂ Gmix =0 (2.37) ∂vB2 generates the spinodal. The region between the binodal and spinodal curves (Region II) represents the region of two-phase metastability. Below the spinodal (Region III), the solution is unstable, phase separation occurs, and two liquid solutions are formed whose relative amounts are given by the lever rule. Note that as temperature increases, the solvating power of the solvent increases, and a single polymer–solvent phase becomes stable.

196

THERMODYNAMICS OF CONDENSED PHASES

two liquid phase region

T

Tc(lower) single phase region

Tc(upper)

vA

Figure 2.34 Schematic diagram of two phase regions resulting in UCST (bottom) and LCST (top). Reprinted, by permission, from J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed., p. 175: Copyright  1991 by Chapman & Hall.

We can substitute our expression for the free energy of mixing in Eq. (2.78) into the stability criteria of (2.36) and (2.37) to solve for the critical interaction parameter at the onset of phase separation, χc χc =

1 2 2vA,c

(2.79)

where vA,c is the volume fraction of the solvent at the critical point. The instance we have considered here, that of a polymer in a poor solvent, results in an upper critical solution temperature (UCST) as shown in Figure 2.33. This occurs due to (a) decreased attractive forces between like molecules at higher temperatures and (b) increased solubility. For some systems, however, a decrease in solubility can occur, and the corresponding critical temperature is located at the minimum of the miscibility curve, resulting in a lower critical solution temperature (LCST). This situation is illustrated in Figure 2.34. 2.3.2

Cohesive Energy Density∗

Solvent–polymer compatibility problems are often encountered in industry, such as in the selection of gaskets or hoses for the transportation of solvents. A rough guide exists to aid the selection of solvents for a polymer, or to assess the extent of polymer–liquid interactions. A semi empirical approach has been developed by Hildebrand based on the principle of “like dissolves like.” The treatment involves relating the enthalpy of mixing to a solubility parameter, δ, and its related quantity, δ 2 , called the cohesive energy density: Hmix = (δA − δB )2 vA vB (2.80) For small molecules, such as solvents, the cohesive energy density is the energy of vaporization per unit volume. The value of the solubility-parameter approach is that

THERMODYNAMICS OF POLYMERS

197

Table 2.6 Typical Values of the Solubility Parameter for Some Common Polymers and Solvents δA [(J/cm3 )1/2 ]

Solvent n-Hexane Carbon tetrachloride Toluene 2-Butanone Benzene Cyclohexanone Styrene Chlorobenzene Acetone Tetrahydrofuran Methanol Water

14.8 17.6 18.3 18.5 18.7 19.0 19.0 19.4 19.9 20.3 29.7 47.9

Polymer

δB [(J/cm3 )1/2 ]

Polytetrafluoroethylene Poly(dimethyl siloxane) Polyethylene Polypropylene Polybutadiene Polystyrene Poly(methyl methacrylate) Poly(vinyl chloride) Poly(vinyl acetate) Poly(ethylene terephthalate) 66-Nylon Polyacrylonitrile

12.7 14.9 16.2 16.6 17.6 17.6 18.6 19.4 21.7 21.9 27.8 31.5

Source: F. W. Billmeyer, Textbook of Polymer Science, 3rd ed. Copyright  1984 by John Wiley & Sons, Inc.

δ can be estimated for both polymer and solvent. As a first approximation and in the absence of strong interactions such as hydrogen bonding, solubility can be expected if δA − δB is less than 3.5 to 4.0, but not if appreciably larger. A few typical values of δ are given in Table 2.6. Values are readily available in the literature. Values of δ for polymers of known structure can be estimated using the molar-attraction constants, E of Table 2.7:  ρ E (2.81) δB = M0 where ρ is the polymer density, and M0 is the repeat unit molecular weight.

2.3.3

Polymer Blends∗

A polymer blend is created when two miscible polymers are mixed together. As in the case of composites, the impetus for creating polymer blends is to combine attributes of the two polymers to create a new material with improved performance. In practice, this is difficult with polymer–polymer solutions, since most common polymers do not mix well with one another to form homogeneous, one-phase solutions. 2.3.3.1 Flory–Huggins Approach. One explanation of blend behavior lies in the thermodynamics of the preceding section, where instead of a polymer–solvent mixture, we now have a polymer–polymer mixture. In these instances, the heat of mixing for polymer pairs (labeled 1 and 2) tends to be endothermic and can be approximated using the solubility parameter. The interaction parameter for a polymer–polymer mixture, χ12 , can be approximated by V0 (δ1 − δ2 )2 (2.82) χ12 = RT

198

THERMODYNAMICS OF CONDENSED PHASES

Table 2.7

Molar Attraction Constants E[(J-cm3 )1/2 / mol]

Group

E[(J-cm3 )1/2 / mol]

Group

CH3

303

CH2

269

NH

368

CH

176

N

125

C

65

CH2

259

NH2

C

463

725

N

NCO

733

CH

249

C

173

Cl2

701

239

Cl primary

419

aromatic

CH

429

S

C

aromatic

200

Cl secondary

425

O

ether, acetal

235

Cl aromatic

329

O

epoxide

360

F

84

COO

668

Conjugation

47

C

O

538

cis

−14

CHO

599

trans

−28

(CO)2 O

1159

Six-membered ring

OH

462

ortho

−48

350

meta

−103

para

OH aromatic H acidic dimer

−19

−13

−82

Source: F. W. Billmeyer, Textbook of Polymer Science, 3rd ed. Copyright  1984 by John Wiley & Sons, Inc.

where the reference volume, V0 , typically assumes a value of 100 cm3 / mol. The critical value of χ12 at the onset of phase separation, χ12,c , can be estimated from χ12,c

1 = 2



1 1/2

x1

+

1 1/2

x2

2

(2.83)

where xi is the degree of polymerization for each polymer, related to its actual degree of polymerization, xn , and the reference volume by   VR (2.84) xi = xn V0 with VR the molar volume of the repeat unit.

THERMODYNAMICS OF COMPOSITES

199

Table 2.8 Complementary Groups Found in Miscible Polymer Blends Group 1 1.

CH2

Group 2

CH

CH2

CH O

2.

CH2

CH

CH3

R O R

3.

CH2 O

4.

CH2 O

5.

CR C

C

CH2

CH2

CH Cl

OCH3 C O

6.

CF2

OCH3

CR

R1 O

CH2

R2

C

O

CH2

CH Cl

O

CH O

C

ONO2

CH3

O

7.

CH2 O

CH C

O

8.

ONO2

CH3 O

O

CH2

CH2

O

S O

Source: J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed. Copyright  1991 by Chapman & Hall.

The situation will change if the enthalpy of mixing is negative, as this will encourage mixing. The search for miscible polymer blends has focused on combinations in which specific intermolecular interactions, such as hydrogen bonds, dipole–dipole interactions, ion–dipole interactions, or charge transfer complex formation, exist. It is now possible to identify certain groups or repeat units, which when incorporated in polymer chains tend to enter into these intermolecular attractions. A short list of some of these complementary groups is given in Table 2.8, where a polymer containing groups from column 1 will tend to form miscible blends with polymers containing groups from column 2.

200

2.4

THERMODYNAMICS OF CONDENSED PHASES

THERMODYNAMICS OF COMPOSITES

Much of what we need to know about the thermodynamics of composites has been described in the previous sections. For example, if the composite matrix is composed of a metal, ceramic, or polymer, its phase stability behavior will be dictated by the free energy considerations of the preceding sections. Unary, binary, ternary, and even higher-order phase diagrams can be employed as appropriate to describe the phase behavior of both the reinforcement or matrix component of the composite system. At this level of discussion on composites, there is really only one topic that needs some further elaboration: a thermodynamic description of the interphase. As we did back in Chapter 1, we will reserve the term “interphase” for a phase consisting of three-dimensional structure (e.g., with a characteristic thickness) and will use the term “interface” for a two-dimensional surface. Once this topic has been addressed, we will briefly describe how composite phase diagrams differ from those of the metal, ceramic, and polymer constituents that we have studied so far. 2.4.1

Interphase Formation via Adhesion, Cohesion, and Spreading

We will see in subsequent chapters, particularly those concerning mechanical properties, that the physical properties of a composite are dictated to a large degree by the interphase between the matrix and reinforcement phases. The interphase plays an increasingly larger role as its proportion relative to the other constituents increases; for example, the surface-to-volume ratio of the reinforcement increases as its characteristic dimension decreases. As a result, it is important to understand the energetics of interphase formation. This is typically done by dividing the formation process into a combination of three possible processes: adhesion, cohesion, and spreading. Let us consider two hypothetical phases in our composite, A and B, without specifying their physical state. They could be a polymer melt and a glass fiber reinforcement during melt infiltration processing, a metal powder and ceramic powder that are being subjected to consolidation at elevated temperature and pressure, or two immiscible polymer melts that will be co-extruded and solidified into a two-phase, three-dimensional object. In any case, the surface that forms between the two phases is designated AB, and their individual surfaces that are exposed to their own vapor, air, or inert gas (we make no distinction here) are labeled either A or B. The following three processes are defined as these surfaces interact and form: Cohesion Adhesion Spreading (B on A)

No surface → 2A (or 2B) surfaces 1 AB surface → 1 A + 1 B surface 1 A surface → 1 AB + 1 B surface

These processes are schematically illustrated in Figure 2.35. We are interested in the free energy change that accompanies each process. The separation process shown in Figure 2.35a consists of the formation of two new interfaces, each of unit cross-sectional area, at a location where no interface previously existed. The free energy change associated with the separation process comes directly from the definition of surface energy [Eq. (2.61)] where two surfaces of unit surface area are formed. With appropriate assumptions regarding constant temperature, pressure, and incompressible fluids, we can equate this free energy change with the

THERMODYNAMICS OF COMPOSITES

201

A A

Unit Area

∆G = WAA A (a)

A

A

Unit Area

∆G = WAB

B

B (b)

B

Unit Area

Unit Area B

∆G = SB/A A

A (c )

Figure 2.35 Schematic illustrations of the processes for which G equals (a) the work of cohesion, (b) the work of adhesion, and (c) the work of spreading. Reprinted, by permission, from P. Heimenz, Principles of Colloid and Surface Chemistry, 2nd ed., p. 316. Copyright  1986 by Marcel Dekker, Inc.

work required to form the two surface, which is called the work of cohesion and is labeled WAA : (2.85) G = 2γA = WAA Here γA is the surface energy of component A. For the case where A is separated from B, as shown in Figure 2.35b, we once again have the formation of two free surfaces, A and B, but must subtract from the energy associated with their formation the energy that was associated with the interface AB, γAB , in order to obtain the free energy change of this process, which is called adhesion. Again, equating work with the free energy change leads to the work of adhesion, WAB . G = γA + γB − γAB = WAB

(2.86)

The work of adhesion measures the degree of attraction between the two phases. For the specific case where A is a solid (S) and B is a liquid (L), γA becomes the solid surface energy, γS , γB become the liquid surface tension, γL , and Young’s equation, Eq. (2.72),

202

THERMODYNAMICS OF CONDENSED PHASES

can be inserted into Eq. (2.86) to obtain a relationship between the liquid–solid contact angle, θ , liquid surface tension and the work of adhesion: WSL = γL (1 + cos θ )

(2.87)

Finally, for the case of spreading of B on A shown in Figure 2.35c, the free energy change is G = γAB + γB − γA (2.88) As usual with free energies, a negative value means that the process occurs spontaneously. We define a quantity called the spreading coefficient, SB/A , which is the negative of the free energy of spreading, such that a thermodynamically favorable spreading process (negative free energy) leads to a positive value of the spreading coefficient. By combining Eqs. (2.85), (2.86), and (2.88), we note that the spreading coefficient can be related to the work of adhesion and work of cohesion: SB/A = WAB − WBB

(2.89)

If WAB > WBB , the A–B interaction is sufficiently strong to promote the wetting of A by B. Conversely, no wetting occurs if WBB > WAB , since the work required to overcome the attraction between two B molecules is not compensated for by the attraction between A and B. Hence, a negative spreading coefficient means that B will not spread over A.

Cooperative Learning Exercise 2.6

Consider the case of an adhesive joint created by bonding two poly(vinyl chloride) (PVC, component A) sheets together with an epoxy adhesive (component B). The surface energies for the PVC solid surface, liquid epoxy, and PVC/epoxy interface are 47 erg/cm2 , 41.7 erg/cm2 , and 4.0 erg/cm2 , respectively. Person 1: Calculate the work of cohesion within the epoxy adhesive. This is the work required to separate epoxy bonds from one another. Person 2: Calculate the work of adhesion between one surface of the PVC sheet and the liquid epoxy. This is the work required to separate the epoxy from the PVC. Combine your results to calculate the spreading coefficient, SB/A . Will the epoxy wet the PVC? Assuming that complete wetting can occur (and that a sufficiently thin layer of adhesive is applied), how will the adhesive joint fail when stressed; i.e., will the epoxy or the PVC/epoxy interface fail first? Answer : WBB = 2(41.7) = 83.4 ergs/cm2 ; WAB = 47 + 41.7 − 4 = 84.7 erg/cm2 ; SB/A = 84.7 − 83.4 = 1.3 erg/cm2 ; The epoxy will wet the PVC. Since WAB > WBB , cohesive failure will occur prior to adhesive failure, and the epoxy will fail before the interface.

2.4.2

Composite Phase Diagrams∗

Unlike the unary, binary, and ternary phase diagrams of the previous sections, there are no standardized guidelines for presenting phase information in composite systems. This

THERMODYNAMICS OF COMPOSITES

203

is due, in part, to the fact that even when two of the composite components are well mixed, the components are usually easily identified on the macroscopic scale. These macroscopic heterogeneities, while potentially beneficial from a functional standpoint, make it difficult to apply classical thermodynamics, which are based on interactions at the molecular level. Nevertheless, the free energy concept can be applied to composite systems, and phase diagrams can be generated from free energy expressions. For example, polymer–clay composites have been modeled in order to study their thermodynamic stability at various concentrations [1–4]. Clay particles are modeled as rigid disks dispersed in the polymer matrix. The composition of the mixture is characterized by the volume fraction of disks. Such factors as the disk length L, disk diameter D, polymer–disk interaction parameter χ, and disk orientation play important roles in the free energy expression. The free energy of mixing for the polymer–disk mixture has the following terms: Gmix = Gconf + Gster + Gint + Gtransl

(2.90)

where the various free energy terms are associated with conformational losses due to alignment of the disks, steric interactions between disks, nonspecific attractive forces between disks, and translational entropy of the disks, respectively. With appropriate assumptions and simplifications, the following free energy expression can be obtained: Gmix = nd [C + ln φd + σ − (b/νd )(ρ − 1) ln(1 − φd ) + χνd φd ] + np ln φp (2.91) where nd is the number of disks in a unit volume, np is the number of polymer molecules in a unit volume, vd is the volume of an individual disk (πD 2 L/4),φd is the volume fraction of disks in the composite such that the volume fraction of polymer φp is (1 − φd ), σ is the conformational free energy (Gconf ), ρ is a function of the angle between two disks, and b is π22 D 3 /16. As with all the phase diagrams we have generated, the equilibrium conditions are applied; that is, the chemical potentials of any 0.5

Xn = 1 Xn = 10 Xn = 100 Xn = 1000

0.4

c

0.3 0.2 0.1 0.0 −0.1 0.0

n

i 0.5 f

1.0

Figure 2.36 Phase diagram for clay disks and polymers of different degrees of polymerization, xn . Reprinted with permission from Y. Lyatskaya and A. C. Balazs, Macromolecules, Vol. 31, p. 6676. Copyright  1998 by the American Chemical Society.

204

THERMODYNAMICS OF CONDENSED PHASES

two phases are equivalent at equilibrium. The chemical potentials are obtained directly from Eq. (2.91) by differentiation in accordance with Eq. (2.36). An example of the resulting phase diagram for the polymer–clay system is shown in Figure 2.36. Here the phase diagram is presented as a plot of the interaction parameter versus volume fraction of the disks. The diameter of the disks in this particular system is 30 arbitrary units, and their length is 1. The regions of disk–polymer miscibility lie below the boundaries. Within the regions of miscibility, i indicates an isotropic phase, and n indicates a nematic liquid crystalline phase (see Section 1.3.6.3). Other systems that have been studied include Al2 O3 fibers in NiAl matrices [5], reinforced metal–matrix composites [6], Fe–TiC composites [7], and glass-fiber–polypropylene composites [8].

2.5

THERMODYNAMICS OF BIOLOGICS

Recall from Section 1.5 that tissue is composed not only of cells, but of an extracellular matrix, of which proteins are an important constituent. In this Section, we are concerned primarily with the thermodynamics that dictate protein behavior in biological systems. Let us begin by considering the insertion of a synthetic material into a tissue sample—for example, a metal catheter into the skin, or a metal–ceramic composite artificial hip into a human body. What happens? Does the foreign material stay isolated from the body? Obviously, it interacts with the tissue cells in some fashion. It can be argued that in the case of a hip implant, what we would like to see is cell adhesion—cells sticking to the material in an attempt to integrate it into the body. It turns out that one of the extracellular matrix proteins we described in Chapter 1, fibronectin, plays an important role in cell attachment and spreading. In this process, fibronectin spreads on the foreign surface, after which point the cell attaches to the fibronectin via a specific type of interaction called receptor–ligand binding (see Figure 1.94). What we are concerned with here is the process that allows a specific chemical constituent in a cell or a protein, such as fibronectin, called the ligand, to interact, or bind, to the corresponding specific chemical constituent in another cell membrane, called the receptor. In this process, the ligand, which is sort of a chemical “key,” fits into the chemical “lock” of the receptor, which leads to a receptor–ligand complex. Receptor–ligand binding is a type of specific binding; that is, there is one receptor that is intended to interact with one ligand. There are other types of nonspecific binding that can occur and contribute to a small extent to the overall binding of the cell to the protein. Let us examine these two situations independently, by first describing the adhesion and spreading of a cell on a surface and then examining specific cell–cell interactions. The discussion at this point will be purely thermodynamic in nature. The kinetics of the process—that is, how fast or slow the adhesion and spreading occurs—is the subject of Chapter 3. 2.5.1

Cell Adhesion and Spreading on Surfaces

For cells the processes of adhesion and spreading on surfaces, are two distinctly separate processes. To date, only adhesion has been described in a thermodynamic way. Cell spreading, on the other hand, is difficult to describe in this manner, due to cellular activity, such as protein production and cytoskeleton transport which complicate the spreading process. As a result, the strength of adhesion does not correlate well with the

THERMODYNAMICS OF BIOLOGICS

205

amount of spreading the cell has undergone. Let us look first at the thermodynamics of cell adhesion on a substrate, then separately at the spreading process. The free energy change for adhesion of cells from a liquid suspension onto a solid substrate is given by (2.92) Gadh = γcs − γcl − γsl where the interfacial free energies are for cell–solid (γcs ), cell–liquid (γcl ), and solid–liquid (γsl ) interfaces, respectively. If the free energy change of adhesion is negative, the adhesion process is spontaneous. If it is positive, the process is thermodynamically unfavorable. As a result, neither extremely hydrophilic nor extremely hydrophobic (γsl < 40 dyn/cm) substrates promote adhesion of cells such as fibroblasts. This trend, while still the subject of much experimental investigation and verification, is illustrated in Figure 2.37. Here, area A represents a nonadhesive zone between 20 and 30 dyn/cm critical surface energy. Area B represents biomaterials with good adhesive properties. Other factors such as surface charge, surface topography, porosity, and mechanical forces can greatly influence the ability of a cell to adhere to a surface.

Cooperative Learning Exercise 2.7

It has been observed that receptor–ligand binding processes are generally exothermic (H < 0). Protein adsorption on surfaces can be described with thermodynamic arguments similar to those used for cell adhesion—that is, with free energy arguments similar to those in Eqs. (2.12) and (2.92). For example, the enthalpies of adsorption for two proteins, HSA and RNase, on α-Fe2 O3 and negatively charged polystyrene (PS-H), respectively, at various pH levels are given below.

Protein

Substrate

pH

HSA

α-Fe2 O3

RNase

PS-H

5 7 5 11

Enthalpy (mJ/m2 ) +1.9 +7.0 +4 −2

Discuss with your neighbor how protein adsorption can still be spontaneous (G < 0), even though some of these processes are endothermic (H > 0). Look up values for the solid surface energies of the two substrates in Appendix 4 (you may have to select similar materials, but this is sufficient for these purposes). How do these enthalpies compare in magnitude to the respective surface energies? Answer : According to Equation (2.12), the entropy must also be considered. In fact, many of these binding/adhesion processes are entropy-driven–even the exothermic ones. The surface energy of Fe2 O3 is probably similar to FeO–about 500 erg/cm2 , and PS is around 33 erg/cm2 . Thus, the enthalpy of adsorption is much larger relative to the surface energy for the RNase/PS system than for the HSA/Fe2 O3 system. Note that 1 mJ/m2 = 1 erg/cm2 .

THERMODYNAMICS OF CONDENSED PHASES

Relative biological interaction

206

A

0

10

20

B

30

40

50

60

70

80

Critical surface tension (dynes/cm)

Figure 2.37 Relationship between the biological interactiveness and the critical surface tension of a biomaterial. Reprinted, by permission, from J. M. Schakenraad, Cells: Their Surfaces and Interactions with Materials, in Biomaterials Science: An Introduction to Materials in Medicine, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, p. 146. Copyright  1996 by Academic Press.

Once the cell is adhered to the surface, it may undergo spreading. Cell spreading is the combined process of continued adhesion and cytoplasmic contractile meshwork activity. This activity involves the formation of microfilaments and lamellar protrusions. When a cell meets the membrane of another cell, further spreading is prohibited. Due to these complexities in the spreading process, it is difficult to give a purely thermodynamic explanation for cell spreading. Nonetheless, some correlation between the surface energy of the substrate and the ability of a cell to spread upon it has been noted (see Figure 2.38). Here, the presence of preadsorbed proteins (solid line in Figure 2.38) can affect the ability of the cell to spread relative to a surface free of preadsorbed proteins (dashed line in Figure 2.38). 2.5.2

Cell–Cell Adhesion

Consider two cells, capable of binding to one another (see Figure 2.39). The surface of ˚ a cell consists of a 100-A-thick layer called the glycocalyx, which contains the surface receptors for binding. There are R1T and R2T total surface receptors, respectively, where R1 is the ligand for R2 and R2 is the ligand for R1 . Let B represent the number of receptor–ligand complexes acting to bridge one cell to another, called bridges. The number of free receptors within and outside the cell–cell contact area are Ric and Rio , respectively, where i = 1, 2. Assuming a constant number of total surface receptors on each cell, the conservation of receptors for cell i will be RiT = B + Ric + Rio

(2.93)

THERMODYNAMICS OF BIOLOGICS

207

Relative spreading

1.0

0.5

0

0

10

50

100

gs(mJ/m2)

Figure 2.38 Influence of substrate surface energy on cell spreading with preadsorbed serum proteins (solid line) and without (dashed line). Reprinted, by permission, from J. M. Schakenraad Cells: Their Surfaces and Interactions with Materials, in Biomaterials Science: An Introduction to Materials in Medicine, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, p. 144. Copyright  1996 by Academic Press.

Glycocalyx

Free receptors Contact region Separation distance Bonds

Free receptors

Figure 2.39 Schematic illustration of cell–cell and cell-substrate adhesion. Reprinted, by permission, from Biophysical Journal, 45, 1051. Copyright  1984 by the Biophysical Society.

The corresponding receptor densities are niT = RiT /Ai , nb = B/Ac , nic = Ric /Ac , and nio = Rio /(Ai − Ac ), where Ai is the total surface area of cell i and Ac is the contact area. We will assume that the distance between the cell membranes is constant and equal to s. The Gibbs free energy for the system can expressed as G =

 j

(Nj nj µj ) + Ac Ŵ(s)

(2.94)

208

THERMODYNAMICS OF CONDENSED PHASES

where Nj is the number of receptor species (either within, Ric , or outside, Rio , of the contact area or as a bridge, B), nj is the density of each of the species (nic , nio , or nb ), and µj is the chemical potential of each receptor state. The quantity Ŵ(s) is called the nonspecific interaction energy and is the mechanical work that must be done against nonspecific repulsive forces to bring a unit area of membrane from an infinite separation to a cell–cell separation distance of s. The repulsive forces that must be overcome are steric and electrostatic in nature, whereas the attractive forces are of the van der Waals type (see Figure 2.40). At small separation distances, repulsive forces ˚ dominate the cell–cell interaction, and diminish at values of s greater than 200 A. This is important because cell–cell and cell–surface adhesion occur with separation ˚ As a result, repulsive forces dominate, and van der distances between 100 and 300 A. Waals forces are often neglected. So, the nonspecific interaction energy can be modeled empirically as a separation-distance-dependent net repulsion   σ −s Ŵ(s) = exp s δg

(2.95)

where σ is a measure of the glycocalyx stiffness and δg is its thickness. Substitution of (2.95) into (2.94) and expansion of the individual free energy terms of the

10−1 Steric

10−2 10−3

W(J /m2 )

10−4 10−5 Electrostatic 10−6

10−7 Van der waals

0

50

100

150

200

250

300

S (Å)

Figure 2.40 Nonspecific work of adhesion required to separate two like cells as a function of cell–cell separation distances, Reprinted, by permission, from P. Bongrand and G. I. Bell, Cell Surface Dynamics: Concepts and Models, A. S. Perelson, C. DeLisi, and F. W. Wiegel, ed., pp. 459–493. Copyright  1984 by Marcel Dekker.

modified ratio of nonspecific repulsion to specific binding, x

REFERENCES

Region I no adhesion

209

slope = R 2T

Ac = 0 slope = Γ(s) A max R 2T − kBT

Region II stable adhesion 0 < Ac < A max

Region III maximal adhesion

Ac = A max

total number of receptors on cell 1, R1T

Figure 2.41 Predicted phase diagram for cell–cell adhesion. Reprinted, by permission, from D. A. Lauffenburger and J. J. Linderman, Receptors, p. 281. Copyright  1993 by Oxford University Press.

receptor species as described above yields the final form of the free energy change for cell–cell adhesion:   Ac σ −s G = Ric nic µic + Rio nio µio + Bnb µb + (2.96) exp s δg As with all free energy expressions in this chapter—minimization with respect to concentration, in this case—the concentration of receptor sites provides the information for the phase diagram, an example of which is shown in Figure 2.41. There are three regions for cell–cell adhesion in this diagram, according to the values R1T , the total number of receptors on cell 1, and a parameter ξ , which is a dimensionless quantity that characterizes a modified ratio of nonspecific binding repulsion to binding. In region I, no stable adhesive interaction occurs between the cells. The attractive forces offered by the receptor–ligand bonds are not strong enough to overcome the repulsive forces resulting from cell–cell contact. In region II, stable cell–cell adhesion exists with a contact area less than Amax . Here, a balance exists between attractive and repulsive forces. In region III, stable cell–cell adhesion results and the contact area is Amax . Here, attractive forces outweigh repulsive forces and the area of contact is limited only by geometric and morphological constraints. REFERENCES Cited References 1. 2. 3. 4. 5.

Lyatskaya, Y., and A. C. Balazs, Macromolecules, 31, 6676 (1998). Ginzburg, V. V., and A. C. Balazs, Macromolecules, 32, 5681 (1999). Ginzburg, V. V., C. Singh, and A. C. Balazs, Macromolecules, 33, 1089 (2000). Huh, J., V. V. Ginzburg, and A. C. Balazs, Macromolecules, 33, 8085 (2000). Hu, W., W. Wunderlich, and G. Gottstein, Acta Mater., 44(6), 2383 (1996).

210

THERMODYNAMICS OF CONDENSED PHASES

6. Asthana, R., P. K. Rohatgi, and S. N. Tewari, Proc. Adv. Mater., 2, 1 (1992). 7. Popov, A., and M. M. Gasik, Scripta Mater., 35(5), 629 (1996). 8. Avalos, F., M. A. Lopez-Manchado, and M. Arroyo, Polymer, 39(24), 6173 (1998).

Thermodynamics—General Kyle, B. G. Chemical and Process Thermodynamics, 3rd ed., Prentice-Hall, Upper Saddle River, NJ, 1999. Smith, J. M., van Ness, H. C., and M. Abbott, Introduction to Chemical Engineering Thermodynamics, 6th ed., McGraw-Hill, New York, 2001. IUPAC Compendium of Chemical Terminology www.iupac.org/publications/compendium/index.html

Thermodynamics of Materials Gaskell, D. R., Introduction to the Thermodynamics of Materials, 3rd ed., Taylor & Francis, Washington, D.C., 1995. http://www.tms.org/pubs/journals/JOM/9712/Kattner-9712.html

Thermodynamics of Metals and Alloys Gaskell, D. R., Introduction to Metallurgical Thermodynamics, McGraw-Hill, New York, 1981. Devereux, O. F., Topics in Metallurgical Thermodynamics, John Wiley & Sons, New York, 1983. Rao, Y. K., Stoichiometry and Thermodynamics of Metallurgical Processes, Cambridge University Press, Cambridge, 1985. http://cyberbuzz.gatech.edu/asm tms/phase diagrams/

Thermodynamics of Ceramics and Glasses Kingery, W. D., H. K. Bowen and D. R. Uhlmann, Introduction to Ceramics, John Wiley & Sons, New York, 1976. Scholze, H., Glass, Springer-Verlag, New York, 1991. http://www.ceramics.nist.gov/webbook/glossary/ped/glossary.htm

Thermodynamics of Polymers Cowie, J. M. G., Polymers: Chemistry & Physics of Modern Materials, 2nd ed., Chapman & Hall, New York, 1991. Polymer Blends, D. R. Paul and C. B. Bucknall, eds., Vols. 1–2, Wiley-Interscience, New York, 2000. Strobl, G., The Physics of Polymers, 2nd ed., Springer, Heidelberg, 1997.

Thermodynamics of Biologics Matthews, F. L., and R. D. Rawlings, Composite Materials: Engineering and Science, CRC Press, Boca Raton, FL, 1999. Jou, D., and J. E. Llebot, Introduction to the Thermodynamics of Biological Processes, PrenticeHall, Englewood Cliffs, NJ, 1990.

PROBLEMS

211

Thermodynamic Data for Biochemistry and Biotechnology, H. J. Hintz, ed., Springer-Verlag, Berlin, 1986. Molecular Dynamics: Applications in Molecular Biology, J. M. Goodfellow, ed., CRC Press, Boca Raton, FL, 1990. Lauffenburger, D. A. and J. J. Linderman, Receptors, Oxford University Press, New York, 1993.

Computer Simulations Rowley, R., Statistical Mechanics for Thermophysical Property Calculations, Prentice-Hall, Englewood Cliffs, NJ, 1994. http://www.calphad.org

PROBLEMS Level I

2.I.1 Draw schematic Gmix versus XNi curves for the Cu-Ni system shown in Figure 2.4 at 1455, 1300, 1100, and 300◦ C. 2.I.2 Compare the relative efficiencies of H2 and CO as reducing agents for metal oxides; for example, if cobalt oxide, CoO, were to be reduced to the pure metal, which would do a better job and at what temperatures? 2.I.3 At 200◦ C, a 50:50 Pb-Sn alloys exists as two phases, a lead-rich solid and a tin-rich liquid. Calculate the degrees of freedom for this alloy and comment on its practical significance. 2.I.4 For 1 kg of eutectoid steel at room temperature, calculate the amount of each phase (α and Fe3 C) present. 2.I.5 Calculate the amount of each phase present in 1 kg of a 50:50 Ni-Cu alloys at a) 1400◦ C, b) 1300◦ C, and c) 1200◦ C. 2.I.6 Calculate the amount of each phase present in 50 kg of free-cutting brass. Neglect the presence of alloying elements less than 3% by weight. 2.I.7 Schematically sketch and label the resulting microstructure of an austenite phase containing 0.65 wt% C that is cooled to below 727◦ C. Level II

2.II.1 We will see in Chapter 5 that the strength of a material, σ , is the work required to separate surfaces divided by the separation distance. In this way, the cohesive strength, σc , is the work of cohesion, WBB , divided by some characteristic intermolecular distance, xBB . Similarly, the adhesive strength, σa , is the work of adhesion, WAB , divided by some characteristic intermolecular distance, xAB . ˚ for both the epoxy adhesive and the (a) Assuming that xBB = xAB ≈ 5.0 A PVC–epoxy interface, calculate the cohesive strength and adhesive strength of the PVC–epoxy composite joint in Cooperative Learning Exercise 2.6. Which has a higher strength, the epoxy with itself or the PVC–epoxy interface? Recall that 1 erg/cm2 = 1 dyn/cm. (b) Now assume that the thickness of the interface,

212

THERMODYNAMICS OF CONDENSED PHASES

xAB , is increased. At what value of xAB does mechanical failure switch from cohesive to adhesive? What does this tell you about the optimal thickness of an adhesive layer if it is to be stronger than the materials it binds together? 2.II.2 Consider 1 kg of a brass with composition 30 wt% Zn. (a) Upon cooling, at what temperature would the first solid appear? (b) What is the first solid phase to appear, and what is its composition? (c) At what temperature will the alloy completely solidify? (d) Over what temperature range will the microstructure be completely in the α-phase? 2.II.3 You have supplies of kaolin, silica, and mullite as raw materials. Using kaolin plus either silica or mullite, calculate the batch composition (in weight percent) necessary to produce a final microstructure that is equimolar in silica and mullite. 2.II.4 Hot isostatic pressing (HIP), in which an inert gas is applied at high pressure, can be used to consolidate and densify ceramic components. The driving force for pore shrinkage in such cases is the sum of the applied pressure, Pe , and the surface energy-driven component, as given by Eq. 2.69. Calculate the ratio of driving forces for HIPing at 35 MPa compared to conventional sintering. Assume a pore diameter of 5 µm and γs = 700 mN/m at the densification temperature. 2.II.5 The melting points of a series of poly(α-olefin) crystals were studied. All of the polymers were isotactic and had chains substituents of different bulkiness. The results are listed below. Use Eq. (2.12) to derive a relationship between the melting point, Tm , and the enthalpy and entropy of fusion, Hf and Sf , respectively. Use this relationship, plus what you know about polymer crystallinity and structure from Chapter 1, to rationalize the trend in melting point. Tm (◦ C)

Substituent CH3

165

CH2CH3

125 75

CH2CH2CH3 CH2CH2CH2CH3 CH2CH

CH2

CH3

−55 196

CH3 CH3 CH2

C

CH2

CH3

350

CH3

2.II.6 The diagram shown below qualitatively describes the phase relationships for Nylon-66, water, and phenol at T > 70◦ C. Use this diagram to draw arrows that trace the following procedures: (1) Water is added to solution A until the system separates into solution B and C; (2) solution C is removed and water

PROBLEMS

213

is added to solution B until the system separates into D and E; (3) phenol is added to solution C until a solution A′ results which is close to A in concentration. Would you expect solutions C and E, or B and D, to contain the higher-molecular-weight polymer? Briefly explain. Outline a strategy for Nylon fractionation based on steps 1–3. Some steps may be repeated as necessary. Nylon

E C A*

B

D

Phenol

Water

2.II.7 The following data were obtained for the work of adhesion between the listed liquids and carbon black. Use these data, together with the surface tensions of pure liquids from Appendix 4 (or from additional sources as necessary), to calculate the spreading coefficients for the various liquids on carbon black. Liquid Benzene Toluene CCl4 CS2 Ethyl ether H2 O

WAB (erg/cm2 ) 109.3 110.2 112.4 122.1 76.4 126.8

Level III

2.III.1 The Maxwell relations in thermodynamics are obtained by treating a thermodynamic relation as an exact differential equation. Exact differential equations are of the form     dz dz dz = dx + dy = Mdx + N dy dx y dy x where z = z(x, y) and M and N correspond to the partial derivatives of z with respect to x and y, respectively. A useful relation in exact differential equations is that     ∂M ∂N = ∂y x ∂x y (Can you prove why this is so?) When x, y, and z are thermodynamic quantities, such as free energy, volume, temperature, or enthalpy, the relationship between the partial differentials of M and N as described above are called Maxwell relations. Use Maxwell relations to derive the Laplace equation for a

214

THERMODYNAMICS OF CONDENSED PHASES

sphere at constant temperature using the following thermodynamic expression for free energy, G: dG = V dP − SdT + γ dA where V is the volume of the sphere, P is the pressure, T is temperature, γ is the surface energy, and A is the sphere surface area. 2.III.2

Why does pore coalescence in a glassy material cause an increase in porosity and negative densification? (Hint: Consider two pores containing n moles of gas that become one pore during isothermal coalescence. The driving force for pore shrinkage is 2γ /r. Use this information with the ideal gas law to calculate the change in pore radius.)

CHAPTER 3

Kinetic Processes in Materials

3.0

INTRODUCTION AND OBJECTIVES

Thermodynamics tells us whether or not a process is favorable; kinetics tells us how quickly that process will take place. The rate at which chemical and biological processes occur is of tantamount importance in many instances. For example, we all know that an exposed piece of iron will rust (oxidize) over time, but is this important if the piece will only be used once in the near future? There are three categories of kinetic processes that are of concern to us: the rate at which materials are formed, the rate at which they are transformed, and the rate at which they decompose. In general, formation and decomposition are chemical processes, involving the reaction of two or more chemical species. Transformations, on the other hand, are usually physical processes, such as the melting of ice, and do not involve chemical reaction. We can also have situations in which both physical and chemical kinetic processes occur simultaneously. In all cases, our goal is to be able to describe the rate at which the process is occurring. There are two important concepts that help us accomplish our task: the thermodynamic concept of free energy, G; and a related quantity called the activation energy, Ea . We will introduce these concepts and then see how one, or both, of them can be modified to help determine the rate of processes in the formation, transformation, and decomposition of different materials. By the end of this chapter, you should be able to: ž

ž ž

ž ž ž ž ž ž

Apply the law of mass action to a reaction to determine the equilibrium constant at a specified temperature. Calculate the activation energy and preexponential factor from kinetic data. Identify the dimensionality and calculate rates of a nucleation and growth process from kinetic data. Distinguish between the different types of corrosion in all material classes. Calculate the EMF of a galvanic cell from the half-cell potentials. Use the Nernst equation to calculate the potential of a half-cell not at unit activity. Distinguish between activation and concentration polarization. Distinguish between homogeneous and heterogeneous nucleation processes. Distinguish between stepwise and radical chain polymerizations.

An Introduction to Materials Engineering and Science: For Chemical and Materials Engineers, by Brian S. Mitchell ISBN 0-471-43623-2 Copyright  2004 John Wiley & Sons, Inc.

215

216 ž ž ž ž ž ž

KINETIC PROCESSES IN MATERIALS

Calculate the degree of polymerization, number-average, and weight-average molecular weights given kinetic data for a polymerization reaction. Identify the different types of copolymers, and use the copolymer equation with appropriate kinetic data to determine which type of copolymer will form. Distinguish between the different types of polymerization processes, and identify the equipment and kinetics that are specific to each. Distinguish between the different types of polymer degradation. Distinguish between the different types of vapor deposition processes, and calculate the rate of deposition for each from kinetic data. Determine the rate of receptor–ligand binding from kinetic data.

3.0.1

The Law of Mass Action

We saw in Chapter 2 that an important thermodynamic quantity is the Gibbs free energy, G. The specific functional relationship we use to describe the free energy will depend on whether we are studying a physical or a chemical transformation. For physical processes, such as phase transformations, the most useful form of the Gibbs free energy is its definition given in Chapter 2: G = H − T S

(2.12)

In the case of reacting systems, however, the number of moles is not constant, and it is more convenient to consider the total free energy as a sum of individual free energy contributions. For this case, let us begin by considering the generalized reaction shown below, in which reactants A and B go to products C and D: aA + bB

cC + dD

(3.1)

where a, b, c, and d are the stoichiometric coefficients of species A, B, C, and D, respectively, and can be represented generically by νi . Recall that stoichiometric coefficients are positive for products and negative for reactants. There can, of course, be more reactants and/or products than those listed here, and the reaction may be reversible. The Law of Mass Action states that the velocity of the reaction at a given temperature is proportional to the product of the active masses of the reacting substances. In other words, the forward reaction rate, r1 , is a function of the concentrations of A and B raised to their corresponding exponents: r1 = k1 [A]a [B]b

(3.2)

Similarly, the reverse reaction rate is related to the amount of products, C and D, present r2 = k2 [C]c [D]d (3.3) In Eqs. (3.2) and (3.3), k1 and k2 are proportionality constants, or rate constants, for the forward and reverse reactions, respectively, and have units of inverse time (typically seconds) and an appropriate inverse concentration, depending on the manner in which the concentration of chemical species are represented and the magnitude of the stoichiometric coefficients.

INTRODUCTION AND OBJECTIVES

217

At equilibrium, which is a thermodynamic state for which G = 0, the rate of the forward reaction must be equal to the rate of the reverse reaction, r1 = r2 , and we can equate Eqs. (3.2) and (3.3) to give: [A]a [B]b k1 = =K k2 [C]c [D]d

(3.4)

The constant K is the equilibrium constant for the reaction at constant temperature. We can relate the equilibrium constant, K, to the free energy of the system. Recall that the free energy change for a reaction must be less than zero for it to proceed spontaneously, and any system whose overall free energy change is zero is in equilibrium. The free energy of a system is simply the sum of the free energy contributions of each of the components, which are measured using their chemical potentials, µi [see Eq. (2.13) for the definition of chemical potential], and the stoichiometric coefficients  G= νi µi (3.5)

When all the components are in their standard states (see Section 2.1.2), the standard Gibbs free energy, G0 , is given by:  G0 = νi µ0i (3.6) The chemical potentials in Eq. (3.5), µi , and the standard chemical potentials in Eq. (3.6), µ0i , are related to each other by the activity of each species, ai : µi = µ0i + RT ln ai

(3.7)

where R is the gas constant and T is the absolute temperature. The activity of a species, then, is an indication of how far away from standard state that species is. If we subtract Eq. (3.6) from Eq. (3.5) and substitute Eq. (3.7) into the result, we obtain a very useful relationship between the activity of the reactants and products and the free energy for a reacting system:     ν 0 i G − G = RT (3.8) νi ln ai = RT ln ai i

At equilibrium, then, G = 0, and the activities of the products and reactants, ai , are related to their respective concentrations:   [A]a [B]b G0 = −RT ln (3.9) [C]c [D]d and we can identify the term in the logarithm from Eq. (3.4) as the equilibrium constant, K: G0 = −RT ln K (3.10) Rearranging Eq. (3.10) to express K in terms of the standard free energy change gives   −G0 K = exp (3.11) RT

218

KINETIC PROCESSES IN MATERIALS

Equation (3.11) gives the effect of temperature on the equilibrium constant. Note that a plot of ln K versus 1/T should give a straight line of slope −G0 /R. 3.0.2

The Activation Energy

Since the equilibrium constant, K, is a ratio of the forward to reverse reaction rate constants, k1 and k2 , respectively, it is reasonable to assume that the forward and reverse rate constants also follow a relationship similar in form to Eq. (3.11). If we replace the standard free energy with some general energy term for activation, Ea , and include a proportionality constant called the preexponential factor, k0 , we get what is called an Arrhenius expression for the rate constant: k = k0 exp(−Ea /RT )

(3.12)

We will not prove the Arrhenius relationship here, but it falls out nicely from statistical thermodynamics by considering that all molecules in a reaction must overcome an activation energy before they react and form products. The Boltzmann distribution tells us that the fraction of molecules with the required energy is given by exp(−Ea /RT ), which leads to the functional dependence shown in Eq. (3.12). Many of the processes we will study in this and other chapters have parameters that appear to follow the very general Arrhenius exponential relationship. The quantity Ea is called the activation energy and represents the energy barrier for the reaction under consideration. As shown in Figure 3.1, the reactants can be thought of as having to overcome an activation energy barrier of height Ea in order to achieve the activated complex state and proceed to products. The heat of reaction, H , can be either negative (exothermic reaction) or positive (endothermic reaction), in which case the products C and D actually have a higher final energy than the reactants. Let us now see how a wide

Total energy of system E

Ea A+B Reactants ∆H

C+D Products

Reaction coordinate

Figure 3.1 Activation energy barrier for a chemical reaction.

KINETIC PROCESSES IN METALS AND ALLOYS

219

variety of kinetic processes in material formation, transformation, and degradation can be described using these relationships.

3.1

KINETIC PROCESSES IN METALS AND ALLOYS

We begin by examining the kinetics processes of formation reactions in metals and alloys. Although many of these substances are elements or binary mixtures of elements, they can be formed through complex reactions, such as the reduction of metal oxides. We will first examine the formation of an intermetallic compound. 3.1.1

Kinetics of Intermetallic Formation

We made a distinction between alloys and intermetallics in Chapter 1, indicating that intermetallics are more like compounds than alloys (which are more like mixtures), even though both contain two or more metallic components. Nonetheless, both are composed entirely of metallic elements, and a discussion of the formation of either would be appropriate here. We focus on the formation of an intermetallic, TiAl3 , to illustrate how kinetic parameters can be obtained from experimental observations. Consider the dissolution of elemental titanium, Ti, in molten aluminum, Al, at various temperatures that lead to the formation of intermetallic TiAl3 layers. Figure 3.2 shows the experimental data for the decrease in thickness of the titanium as a function of time, and Figure 3.3 shows the increase in thickness of the titanium aluminide layer with time. From these plots, the rate of dissolution and rate of formation can be determined, respectively, for each temperature. These values are presented in Cooperative Learning Exercise 3.1. Application of the Arrhenius equation (3.12) in both instances

Total decrease in thickness of Ti (mm)

90



C

5

4

C 0°

85

3

C 0°

80

2

750 1

°C

700 °C

0

1

2

3

4

5 6 Time (hr)

7

8

9

10

Figure 3.2 Rate of dissolution of Ti at various temperatures.

220

KINETIC PROCESSES IN MATERIALS

12 11

0° C

0° C

80 0° C

9

85

90

Total thickness of TiAl3 layers (mm)

10

8 7



6

C

75

5 4 3

700

°C

2 1 0

1

2

3

4

5

6

7

8

9

10

Time (hr)

Figure 3.3 Rate of formation of TiAl3 at various temperatures.

Cooperative Learning Exercise 3.1

Consider the following data for the dissolution of Ti in molten aluminum to form the intermetallic TiAl3 . Temperature (◦ C) 700 750 800 850 900

Rate of Ti Dissolution, kT i (cm/s)

Rate of TiAl3 Formation, kT iAl3 (cm/s)

1.51 × 10−6 2.605 × 10−6 4.17 × 10−6 6.13 × 10−6 9.36 × 10−6

1.042 × 10−5 1.98 × 10−5 2.777 × 10−5 3.798 × 10−5 6.805 × 10−5

Person 1: Determine the activation energy, Ea,T i , and preexponential factor, k0,T i , for the rate of dissolution of Ti. Person 2: Determine the activation energy, Ea,T iAl3 , and preexponential factor, k0,T iAl3 , for the rate of growth of TiAl3 . Group: Compare your results and suggest a reason for your findings. What other processes could affect the rate of either of these reactions? Answers: Ea,T i = 86 kJ/mol; k0,T i = 0.06 Ea,T iAl3 = 84 kJ/mol; k0,T iAl3 = 5.3

KINETIC PROCESSES IN METALS AND ALLOYS

221

allows us to calculate the activation energy for both processes by plotting ln ki versus 1/T (don’t forget that T is in degrees Kelvin!), with the slope of each line equal to −Ea /R. The fact that the two activation energies calculated in Cooperative Learning Exercise 3.1 are essentially equivalent tells us that the rate of the chemical reaction between these two metals, and not the diffusion of either aluminum or titanium through the intermetallic layer formed on the surface of the solid titanium determines the overall rate of TiAl3 formation. As we will see, diffusion is also an activated process and can be described using an Arrhenius-type relationship. 3.1.2

Kinetics of Phase Transformations in Metals and Alloys

3.1.2.1 Nucleation and Growth (Round 1). Phase transformations, such as the solidification of a solid from a liquid phase, or the transformation of one solid crystal form to another (remember allotropy?), are important for many industrial processes. We have investigated the thermodynamics that lead to phase stability and the establishment of equilibrium between phases in Chapter 2, but we now turn our attention toward determining what factors influence the rate at which transformations occur. In this section, we will simply look at the phase transformation kinetics from an overall rate standpoint. In Section 3.2.1, we will look at the fundamental principles involved in creating ordered, solid particles from a disordered, solid phase, termed crystallization or devitrification. Consider the isothermal transformation of a disordered (amorphous) solid to an ordered (crystalline) solid. Obviously, there must be enough thermal energy to allow individual atoms to move around and reorient themselves, but assuming this is possible, it is generally found that at constant temperature, the amount of amorphous material transformed to crystalline material, dx (on a volume basis) per unit time, dt, is given by

dx = nk(1 − x)t n−1 dt

(3.13)

where the quantity (1 − x) is the fraction of amorphous material remaining, n is the reaction order, t is time, and k is the reaction rate constant as expressed by Eq. (3.12). The solution to this equation for the amount of amorphous materials transformed, x, at time t, is known as the Johnson–Mehl–Avrami (JMA) equation [1–3]: x = 1 − exp[−kt n ]

(3.14)

Taking the natural log of this equation twice yields a useful form of the JMA equation: ln[− ln(1 − x)] = ln k + n ln t

(3.15)

Thus, a plot of ln[− ln(1 − x)] versus ln t yields a straight line of slope n. The value of n is important not only because it indicates the reaction order, but also because it has some physical interpretation, depending on whether the crystals are growing at the surface of a material or in the bulk (sometimes called volume crystallization). See Table 3.1 for a description of the dimension of the crystals for various values of n. The intercept of this plot gives the rate constant, k. Obtaining k at different

222

KINETIC PROCESSES IN MATERIALS

Table 3.1 Values of the Growth Dimension, n (a.k.a Reaction Order), for Different Crystallization Processes Mechanism

n

Bulk nucleation, 3-dimensional growth Bulk nucleation, 2-dimensional growth Bulk nucleation, 1-dimensional growth Surface nucleation

4 3 2 1

isothermal temperatures and applying Eq. (3.12) gives the activation energy for the amorphous–crystalline transformation. Most industrially relevant transformation processes are not isothermal; and even in a controlled laboratory environment, it is difficult to perform experiments that are completely isothermal. The kinetics of nonisothermal phase transformations are more complex, of course, but there are some useful relationships that have been developed that allow for the evaluation of kinetic parameters under nonisothermal conditions. One such equation takes into account the heating rate, φ usually in K/min, used in the experiment [4]: ln[− ln(1 − x)] = n ln φ + ln k + constant (3.16) An equation developed by Kissinger and modified by others can be used to determine the activation energy using the temperature at which the transformation rate is a maximum, Tp , for various heating rates [5]:   Ea φ = constant − (3.17) ln 2 Tp RTp The temperature of maximum transformation rate is easily determined using either of two similar techniques called differential scanning calorimetry (DSC) or differential thermal analysis (DTA). These techniques are extremely useful in the kinetic study of both isothermal and nonisothermal phase transformations. 3.1.2.2 Martensitic Transformations. Some phase transformations in the solid phase are so rapid that they approach the speed of sound. These congruent phase transformations (no change in composition) that involve the displacement of atoms over short distances, rather than diffusion (such as in the previous section), are termed displasive or martensitic transformations. The term “martensitic” comes from this type of transformation that is observed in the cooling of austenitic steel (see Section 2.1.3). Martensite is a supersaturated solution of carbon in ferrite with interstitial carbon atoms in the body-centered tetragonal structure that forms when face-centered cubic iron with carbon, called austenite, is rapidly cooled to about 260◦ C. The amount of austenite transformed to martensite is a function of the temperature only, varying from 0 to 100% between the temperatures Ms and Mf , as indicated in the accompanying time–temperature–transformation (TTT) curve shown in Figure 3.4. Martensitic transformations occur in a variety of nonferrous alloy systems, particularly for allotropes that transform at low temperatures. The reasons for this are

KINETIC PROCESSES IN METALS AND ALLOYS

223

800

TE

γ Two-phase product is called pearlite

600

T (°C)

γ+ 400

α+

Sta

200

ish Fe

Two-phase product is called bainite

3C

rt

Ms

α + Fe3C

Fin

γ+M

M + bainite

γ + M + Bainite

Mf M 0 1

10

102

103

104

105

106

t (sec)

Figure 3.4 TTT diagram for eutectoid steel composition (0.80 wt% C). From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

that (a) the free energy difference between the high-temperature and low-temperature phases becomes more negative with decreasing temperature, and (b) the crystal structures of allotropes are relatively simple and similar to one another. Examples of some common martensitic transformations are given in Table 3.2. The martensitic transformation plays an important role in a special class of alloys called shape memory alloys. These materials have the unique ability to “remember” their dimensions before deformation, and they return to their original shape upon undergoing the martensitic transition. This process is shown schematically in Figure 3.5 for the one-way shape memory effect (SME). There is also a two-way shape memory effect that we will not describe here. The shape memory effect is the result of a partial ordering of the structure as the BCC to CsCl structure transition takes place (see Figure 3.6a and3.6b). The martensitic phase (CsCl structure in this case) is easily deformed, and a distorted multidomain martensite phase results (Figure 3.6c). When the object is reheated, the austenitic phase (BCC in this case) reforms and returns to its original structure. The shape memory effect has been studied for applications ranging Table 3.2 tions

Some Common Martensitic Transforma-

Alloy FCC Co HCP Co Cu–Zn–Al αTi βTi Cu–Al–Ni 50% Ni–50% Ti

Transition Temperature (◦ C) 427 −200 to +120 883 −200 to +170 −200 to +100

224

KINETIC PROCESSES IN MATERIALS

Temperature T1 Deformed

T1 Warmed up

T2 > T1 Warmed up Cooled down

T1 One-way

Figure 3.5 Illustration of the one-way shape memory effect.

from drive shafts that stiffen in response to increased shaft vibration [6], to orthodontic wires and other potential biomedical uses [7, 8]. 3.1.3

Kinetics of Corrosion in Metals and Alloys

Corrosion is the deterioration of a material by reaction with its environment. Although the term is used primarily in conjunction with the deterioration of metals, the broader definition allows it to be used in conjunction with all types of materials. We will limit the description to corrosion of metals and alloys for the moment and will save the degradation of other types of materials, such as polymers, for a later section. In this section, we will see how corrosion is perhaps the clearest example of the battle between thermodynamics and kinetics for determining the likelihood of a given reaction occurring within a specified time period. We will also see how important this process is from an industrial standpoint. For example, a 1995 study showed that metallic corrosion costs the U.S. economy about $300 billion each year and that 30% of this cost could be prevented by using modern corrosion control techniques [9]. It is important to understand the mechanisms of corrosion before we can attempt to control it. 3.1.3.1 Types of Corrosion. There are many ways in which to categorize corrosion. One system classifies corrosion into two general categories: macroscopic and microscopic (see Figure 3.7). Macroscopic corrosion is basically “outside-in” corrosion and includes such phenomena as pitting and galvanic (two-metal) corrosion. Microscopic corrosion is more of an “inside-out” process, where such environmental influences as mechanical stress can cause intergranular corrosion and eventually cracking deep within the sample. Another way to classify corrosion is according to the type of environment—that is, wet versus dry corrosion. This is a useful classification scheme for describing why ordinary steel is relatively resistant to dry chlorine gas, but moist chlorine gas, and especially hydrochloric acid solutions, can cause severe damage to steel.

KINETIC PROCESSES IN METALS AND ALLOYS

225

(a)

(b) Tension

Compression (c)

Figure 3.6 Shape-memory alloys transform from (a) a partially ordered, high-temperature austenitic phase to (b) a mixed austenite–martensite low-temperature state to (c) an ordered mixed-phase state under deformation.

For our purposes, the classification system is not as important as (a) recognizing that many different forms of corrosion exist and (b) understanding the fundamental kinetic processes behind these different types. To this end, we will first look at the important principles of corrosion and then see how they can be applied to some important types of corrosion.

226

KINETIC PROCESSES IN MATERIALS

Macroscopic Corrosion

Pitting Flow

Uniform

Attack Erosion

Deposit

Flow Attack Crevice

Cavitation Load

Movement

Fretting

Galvanic Coating

Fluid Water Line Attack

Filiform Corrosion

Microscopic Corrosion

Selective Corrosion

Exfoliation

Intergranular

End-Grain Attack

(Specific Environment) Stress

(Environment)

Stress

Stress Corrosion Cracking

Cyclic Stress Corrosion Fatigue

Figure 3.7 Types of corrosion. Reprinted, by permission, from P. Elliott, Chem. Eng. Prog., 94(5), 33. Copyright  1998 by the American Institute of Chemical Engineers.

3.1.3.2 Electrochemical Reactions. Consider a simple galvanic cell, composed of two metal electrodes, zinc and copper, immersed in two different aqueous solutions of unit activity—in this case, 1.0 M ZnSO4 and 1.0 M CuSO4 , respectively, connected by an electrical circuit, and separated by a semipermeable membrane (see Figure 3.8). The membrane allows passage of ions, but not bulk flow of the aqueous solutions from one side of the cell to the other. Electrons are liberated at the anode by the oxidation (increase in the oxidation number) of the zinc electrode:

Zn

Zn2+ + 2e−

(3.18)

KINETIC PROCESSES IN METALS AND ALLOYS

Zinc anode

e−

227

Load Copper cathode

Porous plug

ZnSO4 soln

Figure 3.8

CuSO4 soln

Schematic diagram of an electrochemical cell.

Electrons flow through the circuit connecting the two electrodes, and they are consumed at the cathode by the reduction (decrease in oxidation number) of copper ions in solution: Cu2+ + 2e− Cu (3.19) The overall reaction, then, is the sum of these two half-reactions: Zn + Cu2+

Zn2+ + Cu

(3.20)

The sulfate ions that remain as a result of reaction (3.19) pass from right to left through the semipermeable membrane in order to balance the charges arising from the zinc ions forming at the anode. Current continues to pass through the circuit, and the zinc electrode slowly dissolves while the copper electrode slowly grows through the precipitation of copper. If we place a high-resistance voltmeter between the copper and zinc electrodes, we measure a potential difference of approximately 1.1 V. The voltage we measure is characteristic of the metals we use. As an additional example, unit activity solutions of CuCl2 and AgCl with copper and silver electrodes, respectively, give a potential difference of about 0.45 V. We could continue with this type of measurement for all the different anode–cathode combinations, but the number of galvanic cells needed would be very large. Fortunately, the half-reactions for most metals have been calculated relative to a standard reference electrode, which is arbitrarily selected as the reduction of hydrogen: 2H+ + 2e−

H2 (g)

(3.21)

The hydrogen electromotive force (EMF) is defined as exactly zero, and all other half-cell EMFs are listed relative to the hydrogen electrode at 25◦ C (see Table 3.3). By convention, the half-cell EMFs are listed as reduction reactions with electrons on the left side of the reaction. All reactions must be reversible, of course, and the corresponding oxidation potential is found by reversing the reaction and reversing the sign on the voltage. For our Zn/Cu galvanic cell example, the EMF we predict for the cell is the sum of the two half-cell reactions. From Table 3.3, this should be the EMF for the reduction of copper (0.337 V) plus the EMF for the oxidation of zinc, which we obtain by reversing the reduction reaction and changing the sign (0.763). The sum of these two EMFs is 1.1 V, as measured earlier. It is important to note that the EMF

228

KINETIC PROCESSES IN MATERIALS

potentials do not need to be multiplied by a factor when balancing half-reactions with different oxidation numbers (why?)—for example, a 2+ cathodic reaction with a 1+ anodic reaction. Test this with the Cu/Ag cell example above, where the measured EMF was about 0.45 V, using the data from Table 3.3. Most environmental conditions do not allow for calculation of these potentials at 25◦ C; in this case, temperature effects cannot be easily neglected. We all know that a car battery has less “cranking power” on a cold winter morning than it does on a warm summer day. Temperature effects can be taken into account by returning to the free energy, which already accounts for temperature fluctuations. Equation (3.8) developed at the beginning of this chapter is where we start. ◦

G − G = RT ln



 i

aiνi



Table 3.3 Standard EMF Potentials for Selected Metals

Corrosion More Likely

Corrosion Less Likely

Electrode Reaction Au3+ + 3e− Pt2+ + 2e− Pd2+ + 2e− Hg2+ + 2e− Ag+ + e− Cu+ + e− Cu2+ + 2e− 2H+ + 2e− Pb2+ + 2e− Sn2+ + 2e− Ni2+ + 2e− Co2+ + 2e− Cd2+ + 2e− Fe2+ + 2e− Ga3+ + 3e− Cr3+ + 3e− Zn2+ + 2e− Mn2+ + 2e− Zr4+ + 4e− Ti2+ + 2e− Al3+ + 3e− Mg2+ + 2e− Na+ + e− Ca2+ + 2e− K+ + e− Li+ + e−

Au Pt Pd Hg Ag Cu Cu H2 Pb Sn Ni Co Cd Fe Ga Cr Zn Mn Zr Ti Al Mg Na Ca K Li

Standard Potential, E (volts), 25◦ C 1.50 1.2 0.987 0.854 0.800 0.521 0.337 0.000 −0.126 −0.136 −0.250 −0.277 −0.403 −0.440 −0.53 −0.74 −0.763 −1.18 −1.53 −1.63 −1.66 −2.37 −2.71 −2.87 −2.93 −3.05

Source: B. D., Craig, Fundamental Aspects of Corrosion Films in Corrosion Science, Copyright  1991 by Plenum Publishing.

(3.8)

KINETIC PROCESSES IN METALS AND ALLOYS

229

For electrochemical systems, we do not have pressure–vapor work that is normally associated with gaseous systems, nor do we have significant thermal effects like heats of fusion or formation. The result is that our usual description of free energy using enthalpies and entropies is not the best representation of free energy here. Instead, the work being generated is electrochemical work that can be best represented by the following relationship: G = −nEF (3.22) where n is the number of electrons taking part in the reaction, F is a constant called the Faraday constant (96,500 coulombs/mol), and E is the potential of the half-reaction in volts. The corresponding standard state free energy is ◦



G = −nE F

(3.23)

Substitution of Eqs. (3.22) and (3.23) into Eq. (3.8) results in the well-known Nernst equation:    ν RT ◦ (3.24) ai i E−E = ln nF i The Nernst equation tells us the electrical potential of a half-cell when the reactants are not at unit activity. 3.1.3.3 Exchange Current Density. Let us now return to our electrochemical cell shown in Figure 3.8. This cell is a combination of two half-cells, with the oxidation reaction occurring at the anode and the reduction reaction occurring at the cathode resulting in a net flow of electrons from the anode to the cathode. Equilibrium conditions dictate that the rate of oxidation and reduction, roxid and rred , be equal, where both rates can be obtained from Faraday’s Law :

roxid = rred =

i0 nF

(3.25)

As in Eq. (3.22), F is the Faraday constant, n is the number of electrons taking part in the reaction, but i0 is a new quantity called the exchange current density. These rates have units of mol/cm2 žs, so the exchange current density has units of A/cm2 . Typical values of i0 for some common oxidation and reduction reactions of various metals are shown in Table 3.4. Like reversible potentials, exchange current densities are influenced by temperature, surface roughness, and such factors as the ratio of oxidized and reduced species present in the system. Therefore, they must be determined experimentally. 3.1.3.4 Polarization. The net current flow produced in a cell results in a deviation of each half-cell potential from the equilibrium value listed in Table 3.3. This deviation from equilibrium is termed polarization, the magnitude of which is given the lowercase greek symbol eta, η and is called the overpotential, E –E ◦ . There are two primary types of polarization: activation polarization and concentration polarization. Activation polarization is an activated process and possesses an activation barrier, just as described earlier in Figure 3.1. There are two parts to the activation process, either one of which can determine the rate of the reaction at the electrode.

230

KINETIC PROCESSES IN MATERIALS

Table 3.4

Some Exchange Current Densities

Reaction 2H+ + 2e− H2 H2 2H+ + 2e− + − 2H + 2e H2 H2 2H+ + 2e− 2H+ + 2e− H2 2H+ + 2e− H2 H2 2H+ + 2e− 2H+ + 2e− H2 H2 2H+ + 2e− 2H+ + 2e− H2 H2 2H+ + 2e− O2 + 4H+ + 4e− 2H2 O 2H2 O O2 + 4H+ + 4e− Fe3+ + e− Fe2+ Ni2+ + 2e− Ni

Electrode

Solution

i0 , A/cm2

Al Au Cu Fe Hg Hg Ni Pb Pt Pd Sn Au Pt Pt Ni

2N H2 SO4 1NHCl 0.1NHCl 2N H2 SO4 1NHCl 5NHCl 1NHCl 1NHCl 1NHCl 0.6NHCl 1NHCl 0.1NNaOH 0.1NNaOH

10−10 10−6 2 × 10−7 10−6 2 × 10−12 4 × 10−11 4 × 10−6 2 × 10−13 10−3 2 × 10−4 10−8 5 × 10−13 4 × 10−13 2 × 10−3 10−6

0.5N NiSO4

Source: M. G. Fontana, Corrosion Engineering, p. 457, 3rd ed. Copyright  1986 by McGraw-Hill Book Co.

The slowest step, or rate-determining step, can be either (a) electron transfer at the electrode–solution interface or (b) formation of atoms at the electrode surface. The activation polarization component of the overpotential, ηa , is related to the actual rate of oxidation or reduction, i, and the exchange current density: ηa = ±β log(i/ i0 )

(3.26)

This equation is often called the Tafel equation. Here, β is a constant, representing the expression 2.3RT /αnF , where R, T , n, and F have been defined previously, and α is the symmetry coefficient, which describes the shape of the rate-controlling energy barrier. Equation (3.26) is represented graphically in Figure 3.9, the resulting plot of which is called an Evans diagram. The linear relationship between overpotential and current density is normally valid over a range of about ±50 mV, but linearity is assumed for most Evans diagrams. The reaction rate, as represented by the current density, changes by one order of magnitude for each 100-mV change in overpotential. Note also that at ηa = 0, the rates of oxidation and reduction are equal, and the current density is given by the exchange current density, i = i0 . The second type of polarization, concentration polarization, results from the depletion of ions at the electrode surface as the reaction proceeds. A concentration gradient builds up between the electrode surface and the bulk solution, and the reaction rate is controlled by the rate of diffusion of ions from the bulk to the electrode surface. Hence, the limiting current under concentration polarization, iL , is proportional to the diffusion coefficient for the reacting ion, D (see Section 4.0 and 4.3 for more information on the diffusion coefficient): DnF C0 (3.27) iL = x

KINETIC PROCESSES IN METALS AND ALLOYS

2H

+

+

2e −

+0.2

231

+b

0

2H

i 0 (H2 /H + )

+

+

−0.1

2e



−b

H2

Overvoltage, ha (V)

H

2

+0.1

−0.2

−0.3 0.001

0.01

0.1

1

10

100

1000

Current density (logarithmic scale)

Figure 3.9 Evans diagram showing effect of activation polarization on overpotential for a hydrogen electrode. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, p. 574, 5th ed. Copyright  2000 by John Wiley & Sons, Inc.

Cooperative Learning Exercise 3.2

Iron metal corrodes in acidic solution according to the following reaction: Fe + 2H+

Fe2+ + H2

Given that the rates of oxidation and reduction of the half-reactions are controlled by activation polarization only, that βFe = +0.07 and βH2 = −0.08, and that the exchange current densities for both the oxidation of Fe and reduction of hydrogen in acidic solution are identical, use the data in Tables 3.3 and 3.4 to determine the following quantities. Recall that the potential for each half-cell is the sum of the equilibrium potential and the corresponding overpotential, in this case, ηa . Person 1: Derive an expression for the electrical potential, EH for the reduction of hydrogen as a function of the current density, i. Person 2: Derive an expression for the electrical potential for the oxidation of iron, EFe as a function of the current density, i. Group: Recognize the fact that at equilibrium, EH = EFe and combine your results to determine the current density, i, for the cell, and use this value to calculate the corrosion rate of iron, rFe . Answer : EH = EH+ /H2 + βH log(i/iO,H ); EFe = EFe/Fe2+ + βFe2+ log(i/io,Fe2+ ); i = 1.3 × 10−3 A/cm2 ; rF e = 6.73 × 10−9 mol/cm2 žS

where n and F have been defined previously, C0 is the concentration of reacting ions in the bulk solution, and x is the thickness of the depleted region near the electrode, or

232

KINETIC PROCESSES IN MATERIALS

the distance through which the ions must travel from the bulk to the electrode, termed the Nernst diffusion layer. The overpotential for concentration polarization, ηc , is:   i 2.3RT log 1 − ηc = nF iL

(3.28)

Both activation and concentration polarization typically occur at the same electrode, although activation polarization is predominant at low reaction rates (small current densities) and concentration polarization controls at higher reaction rates (see Figure 3.10). The combined effect of activation and concentration polarization on the current density can be obtained by adding the contributions from each [Eqs. (3.26) and (3.28)], with appropriate signs for a reduction process only to obtain the Butler–Volmer equation: ηreduction = −β log

    i 2.3RT i + log 1 − i0 nF iL

(3.29)

Equation (3.29) allows us to determine the kinetics of most corrosion reactions from only three parameters: β, i0 , and iL . 3.1.3.5 Corrosion Rate. The current densities give us an estimate of the rate of the electrochemical reactions involved in corrosion, but they do not give us a direct measurement of the corrosion rate. A simple, empirical expression can be used to quantify the rate of corrosion in terms of a quantity called the corrosion penetration rate (CPR): KW CPR = (3.30) ρAT

where W is the weight loss of the sample per unit time, t. The variables ρ and A represent the density of the material and the exposed surface area of the sample, respectively. The constant K varies with the system of units being used; K = 534 to +

i0

Overvoltage, h

0 −

Activation polarization

Concentration polarization

iL Log current density, i

Figure 3.10 Effect of activation and concentration polarization on overpotential. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 576. Copyright  2000 by John Wiley & Sons, Inc.

KINETIC PROCESSES IN CERAMICS AND GLASSES

233

give CPR in mils per year (mpy, where 1 mil = 0.001 inch), and K = 87.6 to give CPR in mm/y. 3.2 3.2.1

KINETIC PROCESSES IN CERAMICS AND GLASSES∗ Nucleation and Growth (Round 2)

We described martensitic-type transformations in metals in Section 3.1.2.2, described spinodal decomposition of polymers in Chapter 2, and initiated a description of crystallization processes in Section 3.1.2.1. We wish to elaborate on crystallization by identifying two primary phenomena called nucleation and growth. Nucleation and growth kinetics are composed of two separate processes. Nucleation is the initial formation of small particles of the product phase from the parent phase. The resulting nuclei are often composed of just a few molecules. The growth step involves the increase in size of the nucleated particles. Many natural phenomena follow nucleation and growth kinetics, such as ice formation. Unlike martensitic transformations, both spinodal decomposition and nucleation and growth processes involve diffusion, but there are some subtle differences. Figure 3.11 shows a comparison based on composition and spatial extent between a spinodal transformation and a nucleation and growth transformation. Spinodal transformations are large in spatial extent but involve relatively small concentration differences throughout the sample, whereas the nucleation process involves the formation of small domains of composition very different from the parent phase, which then grow in spatial extent via concentration-gradient-driven diffusion. Let us first consider the liquid–solid phase transformation. At the melting point (or more appropriately, fusion point for a solidification process), liquid and solid are in equilibrium with each other. At equilibrium, we know that the free energy change for the liquid–solid transition must be zero. We can modify Eq. (2.11) for this situation Gv = Hf − Tm Sf = 0

(3.31)

C ′a

Concentration

Co Ca

Early

Later (a)

Final

Early

Later

Final

C ′a Cc Ca

Distance (b)

Figure 3.11 Schematic comparison of dimensional changes that occur in (a) spinodal and (b) nucleation and growth transformation processes. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann. Introduction to Ceramics, Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

234

KINETIC PROCESSES IN MATERIALS

where Gv is the free energy change per unit volume for fusion, Hf is the enthalpy of fusion and is related to the strength of interatomic or intermolecular forces of the solid, Sf is the entropy of fusion, representing the degree of disorder in the solid, and Tm is the melting point. We can solve Eq. (3.31) for the entropy change of fusion: Sf =

Hf Tm

(3.32)

If we assume that the entropy and enthalpy are relatively independent of temperature, we can drop the subscript “m” on temperature in Eq. (3.32), thereby obtaining a more general relationship between the entropy and enthalpy of fusion as a function of temperature. Substitution of this modified form of Eq. (3.32) into Eq. (3.31) gives the following relation for the free energy: Gv = Hf



(Tm − T ) Tm



(3.33)

This expression will be very useful in the discussion of nucleation and growth kinetics. In the nucleation step, there must be sites upon which the crystals can form. This is similar to “seeding” the clouds to cause water to precipitate (rain). There are two sources for these nucleating particles: homogeneous and heterogeneous agents. 3.2.1.1 Homogeneous Nucleation. In homogeneous nucleation, critical nuclei (nuclei with enough molecules to initiate growth) form without the aid of a foreign agent. This happens very rarely in practice, but is useful for theoretical discussions. Consider the formation of a small, solid, spherical particle from the parent (liquid) phase. There is a free energy change, G, associated with this process. Much like the potential energy function between two atoms, there are two competing forces that contribute to the overall free energy change: (1) the free energy change due to a volume change (a negative contribution) and (2) the free energy change due to formation of an interphase (a positive contribution). These forces are quantified in Eq. (3.34):

G =



4πr 3 3



Gv + 4πr 2 γ

(3.34)

where Gv is the free energy change per unit volume for the phase change (which must be negative); r is the nucleated particle radius; γ is the interfacial energy (often called “surface tension” for a liquid–gas interface), which has units of energy/area (erg/cm2 ) or force/length (dyne/cm). The first term in Eq. (3.34) is the free energy change associated with the volume change, and the second term is the free energy change associated with the interphase formation. As with any activated process, there is a free energy barrier, G∗ , or “critical free energy” that must be surmounted in order to nucleate a particle. This is illustrated in Figure 3.12 as a plot of free energy versus the particle size, r. Once the overall free energy change decreases past G∗ , the nucleation process becomes spontaneous. This occurs as enough molecules gather to form a nucleus with critical radius, r ∗ . We can solve for r ∗ by taking the derivative of G in Eq. (3.34) with respect to r, setting

KINETIC PROCESSES IN CERAMICS AND GLASSES

235

(4pr 2)g

∆G ∗ ∆G 0

3 ( 4p 3 r )∆Gv

r∗

0

r

Figure 3.12 Representation of activation energy barrier for nucleation. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

Cooperative Learning Exercise 3.3

Work with a neighbor. Consider Eq. (3.34), which describes the free energy change associated with the homogeneous nucleation of a solid, spherical particle of radius, r, from a parent liquid phase: G =



4πr 3 3



Gv + 4πr 2 γ

Recall that this equation could be minimized with respect to particle radius to determine the critical particle size, r ∗ , as given by Eq. (3.35). This critical radius could then be used to determine the height of the free energy activation energy barrier, G∗ , as given by Eq. (3.36). A similar derivation can be performed for a cubic particle with edge length, a. Person 1: What is the equivalent expression for the first term in Eq. (3.34) for the cube? Recall that Gv is the free energy change per unit volume. Person 2: What is the equivalent expression for the second term in Eq. (3.34) for the cube? Recall that γ is the interfacial energy for the liquid–solid interface. Combine your results to arrive at an equivalent expression for Eq. (3.34) for the cube. Minimize this equation to determine the critical cube length, a ∗ . Substitute this value back into your expression to determine the critical free energy barrier for the cube, Gv ∗ . Why is Gv ∗ higher for a cube than for a sphere? What does this tell you about the likelihood of cubes nucleating rather than spheres? Answer : a 3 Gv ; 6a 2 γ ; a ∗ = −4γ /Gv ; Gv ∗ = 32γ 3 /Gv 2 . The barrier is lower for spherical particles so their formation is more likely.

236

KINETIC PROCESSES IN MATERIALS

it equal to zero (the slope of G vs. r in Figure 3.12 is zero at G = G∗ ), and replacing r by r ∗ to get −2γ (3.35) r∗ = Gv The height of the activation energy barrier, G∗ , can now be found by evaluating Eq. (3.34) at r by r ∗ with the aid of Eq. (3.35) G∗ =

16πγ 3 3G2v

(3.36)

To summarize the nucleation stage, then, particles that achieve a critical radius r > r ∗ may enter the growth stage of this phase transformation process. 3.2.1.2 Heterogeneous Nucleation. In heterogeneous nucleation, the critical nuclei form with the assistance of nucleating sites. These sites can be impurities, container walls, cracks, or pores and may either be naturally present or intentionally added to promote nucleation. The effect of nucleating agents is to decrease G∗ . Unlike homogeneous nucleation, heterogeneous nucleation involves an interface between two compositionally different materials, so we must account for the interaction of the parent phase with the nucleating particle. This is accomplished through the introduction of the contact angle, θ , at the three-phase interphase of the parent liquid, its solid phase, and the nucleating agent (see Figure 3.13 and Section 2.2.2.3). Without derivation, we present the free energy barrier for heterogeneous nucleation, G∗s , as

G∗s = G∗

(2 + cos θ )(1 − cos θ )2 4

(3.37)

where G∗ is the free energy barrier for homogeneous nucleation and the subscript “s” on G∗s indicates heterogeneous or “surface” nucleation. 3.2.1.3

Nucleation Rate. We can now define an overall nucleation rate:  = νns n∗

(3.38)

Liquid gSL Solid

gna− S

q

gna− L

Nucleating agent

Figure 3.13 Schematic diagram of contact angle formed at three phase intersection during heterogeneous nucleation. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

KINETIC PROCESSES IN CERAMICS AND GLASSES

237

where  is the number of critical nuclei that form per unit volume per unit time, ns is the number of molecules in contact with critical nucleus, n∗ is the number of critical size clusters per unit volume, and ν is the collision frequency of single molecules with the nuclei (see Figure 3.14). The number of critical nuclei, n∗ can be related to the free energy barrier, G∗ , through an Arrhenius-type expression: n∗ = n0 exp(−G∗ /kB T )

(3.39)

where n0 is the number of single molecules per unit volume, kB is Boltzmann’s constant, and T is absolute temperature. Similarly, the collision frequency, ν, can be expressed by an Arrhenius relation: ν = ν0 exp(−Gm /kB T )

(3.40)

where ν0 is the molecular jump frequency, and Gm is the activation energy for transport across the nucleus–matrix interface, which is related to short-range diffusion. Putting these expressions back into Eq. (3.38) gives  = νns n0 exp(−G∗ /kB T ) exp(−Gm /kB T )

(3.41)

For heterogeneous nucleation, we simply use G∗s instead of G∗ :  = ν0 ns n0 exp(−G∗s /kB T ) exp(−Gm /kB T )

(3.42)

Though Eqs. (3.41) and (3.42) are somewhat cumbersome, they describe the rate at which nuclei form as a function of temperature. Note that in both equations, ns and n0 have dimensions of number per unit volume, and ν0 has units of time−1 . What would be more useful are forms of Eqs. (3.41) and (3.42) that contain some measurable, physical properties of the system. Free energies hardly fit this description. The danger in doing this is that we have to make some simplifications and assumptions. Let us begin by recalling that the free energy for homogeneous nucleation is given by Eq. (3.36): 16πγ 3 (3.36) G∗ = 3(Gv )2

ns n∗

v

Figure 3.14

Schematic illustration of a nucleation site formation.

238

KINETIC PROCESSES IN MATERIALS

Recall also that for liquid–solid transitions at temperatures close to Tm , Gv is given by Equation (3.33): Hv (Tm − T ) (3.33) Gv = Tm where Hv is the heat of transformation per unit volume. Substitution of Eqs. (3.33) and (3.36) into the first exponential in Eq. (3.42) gives us the temperature dependence of the nucleation rate for homogeneous nucleation:

  −Gm −16πγ 3 T 2 exp  = ν0 ns n0 exp 3kB T Hv2 (Tm − T )2 kB T

(3.43)

While this may appear even more cumbersome than Eq. (3.41), it contains some parameters that are directly measurable such as the interfacial surface energy, γ , and the heat of fusion, Hv , but more importantly, it contains the temperature difference (Tm − T ), which is the degree of undercooling—that is, how far the temperature is below the melting (solidification) point! The two exponentials in Eq. (3.43) compete against each other. As T decreases below the melting point: ž

ž

(Tm − T ) becomes larger (remember that this term is squared, so it is always a positive value), the term inside the first exponential becomes smaller, but because it is negative, the entire exponential gets bigger and the nucleation rate increases due to this exponential—the driving force for nucleation becomes greater. The term in the second exponential becomes more negative, the exponential gets smaller and the nucleation rate decreases due to this exponential term—diffusion becomes more difficult as the temperature decreases and particles cannot migrate to the nucleation surface.

These generalizations are true for both homogeneous and heterogeneous nucleation. As a result, we would expect that the two competing exponentials give rise to a maximum nucleation rate at some temperature below the melting point, and this is indeed the case, as illustrated in Figure 3.15. It is also logical that heterogeneous nucleation has a higher absolute nucleation rate than homogeneous nucleation and that heterogeneous occurs at a higher temperature (lower degree of undercooling) than homogeneous nucleation. 3.2.1.4 Growth. The second portion of nucleation and growth is the growth process. Here the tiny nuclei grow into large crystals through the addition of molecules to the solid phase. There are two primary types of crystal growth: thermally activated (diffusion controlled) and diffusionless (martensitic). We have already discussed martensitic transformations, so only thermally activated crystal growth will be considered here. The development of the proper growth rate expression is highly dependent upon the type of phase transformation—that is, crystallization from the melt, vapor phase, or dilute solution. We will simply use a general form of the growth rate expression which is based upon an Arrhenius-type expression:  = A[1 − exp(G/kB T )]

(3.44)

KINETIC PROCESSES IN CERAMICS AND GLASSES

239

Tm Heterogeneous

T

Homogeneous

0°K

Nucleation Rate

Figure 3.15 Effects of temperature and undercooling on homogeneous and heterogeneous nucleation rates.

where  is the growth rate, G is the molar free energy difference between product and parent phase, and A is a preexponential factor. The form of the preexponential factor, A, depends on the type of theory one wishes to employ. It can be directly related to the liquid phase viscosity, as in A = νa0 , where ν is the frequency factor for transport across the interphase ν=

kB T 3πa 3 η

(3.45)

in which η is the viscosity of the liquid phase, and a0 is the molecular diameter; or A can be related to the diffusivity of atoms across the interphase, as in A = KD, where K is a constant and the diffusivity of atoms jumping across interphase, D, is given by D = D exp



−H 0 kB T



(3.46)

These two expression are actually quite similar since D and η are interrelated (see Section 4.3). Other expressions exist as well for such phenomena as surface growth and screw dislocation growth. 3.2.1.5 Overall Phase Transformation Rate. Finally, the overall transformation rate,  is given by the product of the nucleation and growth expressions [Eqs. (3.41) or (3.42), and (3.44), respectively):  =×

(3.47)

The overall transformation rate is shown qualitatively in Figure 3.16. Notice that the maximum nucleation rate occurs at a lower temperature than the maximum growth rate, and that the maximum transformation rate may not be at either of these two rate maxima. Note also that there is some finite transformation rate, even at very low

240

KINETIC PROCESSES IN MATERIALS

Growth Rate

T0

Overall Transformation

T Nucleation Rate

Rate

Figure 3.16 Schematic representation of transformation rates involved in crystallization by nucleation and growth kinetics.

Melting and forming

Temperature

Growth

Nucleation

Time

Figure 3.17 Schematic time-temperature cycle for the controlled crystallization of a glass-ceramic body.

temperatures. This is why some glasses can crystallize over very long periods of time. As long as there is some molecular motion, there is a probability of crystallization taking place. Finally, it should be pointed out that the breaking down of the liquid to solid phase transformation into two separate steps, nucleation and growth, is not entirely artificial. An interesting application of the distinct nature of the nucleation and growth processes is found in the formation of glass ceramics. As shown in Figure 3.17, certain glassforming inorganic materials can be heat-treated in very controlled ways in order to affect the structure of the crystals that form. For instance, a glass can be rapidly cooled below its melting point, then heated to the maximum nucleation rate, which (recall) is below both the melting point and the maximum growth rate. At this temperature and with sufficient time, many small nuclei will be formed. The glass, now replete with

KINETIC PROCESSES IN CERAMICS AND GLASSES

241

many tiny nuclei, is heated to the maximum growth rate temperature and held there while the crystals grow. The crystals can only grow so large, though, since they soon run into another growing crystal from a neighboring nucleation site. In this way, the size of the crystals, or crystallites, in the glass can be controlled, and a glass ceramic is formed. Glass ceramics are unique in that they are crystalline materials, yet they are transparent in many cases (see Section 1.2.5). They also possess unique physical properties, such as low thermal expansion, due to the large amount of interphase relative to the same material made of larger crystals. In the extreme, crystallites on the order of 10−9 meters in diameter can be formed, resulting in so-called nanostructured ceramics. These materials, like glass ceramics, have unique physical properties, which may one day lead to improved ductility and thermal properties in ceramics. 3.2.2

Kinetics of Corrosion in Ceramics and Glasses

Although the term corrosion is normally associated with metals, the definition presented in Section 3.1.3 is equally applicable to ceramic materials. The cost to industry due to ceramic corrosion is similarly large, over $1 billion annually. For example, the cost to replace the refractory insulation in a glass melting furnace might be about $10 million, nearly 60% of which is related to replacing the insulation itself, with the remainder due to labor costs. As with metallic corrosion, types of ceramic corrosion can be grouped a number of different ways, but classification according to the type of attacking medium (i.e., liquid vs. gaseous corrosion) is a useful distinction. We will investigate these two methods of attack separately. 3.2.2.1 Kinetics of Ceramic and Glass Corrosion by Liquids. The corrosion of a solid, crystalline ceramic material by a liquid occurs through the formation of a reaction product between the solid and the solvent. This reaction product is generally less soluble than the bulk ceramic solid, and it may form an attached surface layer on the solid. This type of corrosion in a liquid is called indirect dissolution, incongruent dissolution, or heterogeneous dissolution. The rate-limiting step in this type of ceramic corrosion can be the interfacial layer formation, diffusion through the interfacial layer, or diffusion through the liquid. Direct dissolution, congruent dissolution, or homogeneous dissolution occurs when the solid, crystalline ceramic material dissolves directly into the liquid through dissociation or reaction. The concentrations of the ceramic species in solution, along with the corresponding diffusion coefficients of the species, determine the rate of direct dissolution. The type of attack that occurs in liquid media is highly dependent on the chemical nature of the liquid—that is, molten metal, molten ceramic, or aqueous solution. We will consider two industrially important cases: attack by molten metals and attack by aqueous media. The attack of most metal oxide ceramics by molten metals involves a simple exchange of one metal ion for another. For example, silicon dioxide in contact with molten aluminum is susceptible to the following corrosion reaction:

3SiO2 (s) + 2Al(l)

2Al2 O3 (s) + 3Si(l)

(3.48)

The ease with which this exchange can occur is given by the free energy of reaction, G, which is the summation of the free energy of formation for the individual oxides,

242

KINETIC PROCESSES IN MATERIALS

presented here at 300 K in kJ/mol: 3SiO2 (s)

3O2 (g) + 3Si(l)

3O2 (g) + 4Al(l) 3SiO2 (s) + 4Al(l)

2Al2 O3 (s) 2Al2 O3 (s) + 3Si(l)

Gf = 2568 Gf = −3164

(3.49)

Gf = −596

Values for formation reactions can be found in the JANAF tables [10]. Another way in which a molten metal can attack an oxide is through compound formation, such as the formation of spinel: 4Al2 O3 (s) + 3Mg(l)

3MgAl2 O4 (s) + 2Al(l)

(3.50)

This reaction has a lower free energy of formation (−218 kJ/mol at 1000◦ C) compared to the simple oxidation–reduction reaction (−118 kJ/mol) analogous to Eq. (3.48) above. It is also possible that the molten metal can completely reduced the metal oxide ceramic, forming an alloy and gaseous oxygen: liquid metal

Mx Oy (s) −−−−−−→ xM(solution) + y2 O2 (g)

(3.51)

Ceramic materials in contact with aqueous liquids are also susceptible to attack. This is of particular importance in the realm of civil engineering, where materials of construction that contain minerals routinely come in contact with groundwater and soil. Minerals dissolve into aqueous solutions through the diffusion of leachable species into a stationary thin film of water, about 110 µm thick Mineral A + nH+ + mH2 O

Mineral B + qM+

(3.52)

where M+ is the soluble species. The equilibrium constant for this equation is given by Eq. (3.4): [M+ ]q K= +n (3.53) [H ] [H2 O]m where the solids species (minerals) have unit activity. In dilute solutions and at one atmosphere pressure, the activity of water is also approximately unity. At higher pressures, the activity of water is approximately proportional to the pressure. It is evident from this relationship that the dissolution of ceramics is highly dependent upon the pH ([H+ ]) of the aqueous medium. Generally, at low pH values (acidic environments) the equilibrium shifts to the right, and corrosion is promoted, whereas high pH values (basic environments) retard corrosion by shifting the equilibrium back to the left. Not all ceramic materials behave the same at a given pH, however. As the material begins to dissolve, ions form at the surface, water molecules orient themselves accordingly, and an electrical double layer is established, as shown in Figure 3.18. The first layer of charged ions and oriented water molecules is called the inner Helmholtz plane (IHP), and the second layer of oppositely charged particles is called the outer Helmholtz

KINETIC PROCESSES IN CERAMICS AND GLASSES

243

IHP OHP

+



+

− +



+

Solution





+

+





+

+



+



+

+







Metal







+



+



+





+

+

+



+



+



+





+

Figure 3.18 Formation of the electrical double layer of a surface in solution, showing the inner Helmholtz plane (IHP) and outer Helmholtz plane (OHP). Reprinted, by permission, from B. D. Craig, Fundamental Aspects of Corrosion Films in Corrosion Science, p. 4. Copyright  1991 by Plenum Press.

plane (OHP). This arrangement should remind you of a capacitor; and indeed, this is how this system is often modeled. The potential established between these two layers decreases as we move away from the surface into the solution. The extent to which this double layer is established and the magnitude of the charge on the surface are directly related to the pH of the surrounding solution. The pH value for which there is no net surface change is called the isoelectric point (IEP). An empirical formula for determining the IEP of oxides has been established [11]: IEP = 18.6 −

11.5z d

(3.54)

where z is the oxidation state of the cation in the film, and d is the distance (in angstroms) of an adsorbed proton from the cation via the oxygen ion. Additional values of IEPs for some typical oxides and hydroxides are given in Table 3.5. Related to the attack of polycrystalline ceramic materials by aqueous media is the hydrolysis of silicate glasses. The following relationship has been developed to describe the effect of time and temperature on the acid corrosion (10% HCl) of silicate

244

KINETIC PROCESSES IN MATERIALS

Table 3.5 Some Isoelectric Points (IEP) for Selected Oxides and Hydroxides Material

IEP

α-Al2 O3 γ -Al2 O3 γ -Al(OH)3 α-Al(OH)3 CuO Fe2 O3 Fe3 O4 CdO Ta2 O5 MO2 MO

5–9.2 8.0 9.25 5.0 9.5 8.6 6.5 12 2.9 0–7.5 8.5–12.5

Material ZnO TiO2 Pb(OH)2 Mg(OH)2 SiO2 ZrO2 SnO2 WO3 M2 O5 , MO3 M 2 O3 M2 O

IEP 8.8 4.7–6.2 11.0 12 2.2 6.7 4.3 0.4 11.5

M = metal cation.

glasses [12]: W = at b1 exp(−b2 /T )

(3.55)

where W is the weight loss, t is time, T is temperature, and a, b1 , and b2 are experimentally determined coefficients. Like aqueous attack of crystalline ceramics, glass dissolution rates are a strong function of pH. Unlike crystalline ceramics, however, there are two competing mechanisms in glass dissolution: ion exchange and matrix dissolution (see Figure 3.19). In ion exchange the metal–oxygen bond is coordinated to hydrogen ions, whereas in matrix dissolution the metal–oxygen bond is coordinated to hydroxyl ions. Ultimately, the corrosion of glasses affects their mechanical properties and determines their useful service lifetime on an industrial scale.

dissolution (mg/ml)

3.2.2.2 Kinetics of Ceramic and Glass Corrosion by Gases. The attack of ceramics and glasses by gaseous reactants is much more prevalent than corrosion due to liquids. In addition to the chemical composition of the attacking medium, the geometry of the ceramic solid is of tantamount importance to its corrosion resistance. That is, a

acid region of ion exchange

neutral region of greater durability

basic region of matrix dissolution

1 2 3 4 5 6 7 8 9 10 11 12 13 pH

Figure 3.19

Effect of pH upon glass dissolution.

KINETIC PROCESSES IN CERAMICS AND GLASSES

245

porous or highly permeable substrate will be much more susceptible to attack by gases than a theoretically dense one. Pore size, pore size distribution, surface-to-volume ratio, and diffusivity all play important roles in this process. Gas or, more appropriately, vapor attack can result in the formation of either a solid, liquid, or gas, depending on the reaction conditions and the type of product formed. For example, Na2 O vapors can attack SiO2 to form liquid sodium silicate. In some cases, the vapor may enter the ceramic due to a thermal gradient, then condense and dissolve the ceramic by liquid attack. The corrosion rate can be limited by any number of steps in the dissolution process: ž ž ž ž ž ž ž ž ž

Diffusion of the gas to the solid. Adsorption of the gas molecule onto the solid surface. Surface diffusion of the adsorbed gas. Decomposition of reactants at surface sites. Reaction at the surface. Removal of products from reaction sites. Surface diffusion of products. Desorption of gas molecules from the surface. Diffusion of products away from the solid.

Notice that most of the possible rate-limiting steps involve movement of atoms—diffusion, adsorption, and desorption—and only one direct chemical reaction is involved. In general, the total pressure of the system will determine which step is ratelimiting. At low total pressures (10−1 atm). We will reserve the majority of our description of diffusion in solids to Chapter 4; but for now, recall that the diffusion coefficient is given by D and has units of cm2 /sec. For gases, the concentration is often represented by the partial pressure of component A, PA , which gives the functional dependence of corrosion rate on pressure, as described previously. When the rate of the chemical reaction occurring at the surface is the rate-limiting step, the principles we have described to this point apply. The reaction rate can have any order, and the gas reacts with the ceramic substrate to produce products. Although our discussion to this point has focused on oxide ceramics, there are a number of nonoxide ceramics, such as carbides, nitrides, or borides, that are of importance and that undergo common decomposition reactions in the presence of oxygen. These ceramics are particularly susceptible to corrosion since they are often used at elevated temperatures in oxidizing and/or corrosive environments. For example, metal nitrides can be oxidized to form oxides: 2Mx O(s) + yN2 (g)

2Mx Ny (s) + O2 (g)

(3.56)

In this case, the reaction rate, r, is a function of the partial pressure of the gaseous species, Pi , since solids are taken to have unit activity: y

r=k

PN2 PO2

(3.57)

246

KINETIC PROCESSES IN MATERIALS

diffusion

solid (I)

solid (II)

gas

Figure 3.20 Diffusion of reaction products through boundary layer.

Note from reaction (3.56) that the chemical composition of the surface is changing. As this surface layer builds over the course of the reaction, diffusion of species back through this surface reaction layer, or boundary layer, can become the rate-limiting step (see Figure 3.20). Again, the intricacies of boundary layer diffusion will be addressed in Chapter 4, but for now, we can look at the kinetics associated with the growth of this boundary layer. For thin reaction films (5 nm), the film thickness follows either a parabolic or rectilinear relationship with time x 2 = k6 t

(3.59a)

x = k7

(3.59b)

In Eqs. (3.58) and (3.59), the ki are the reaction rate constants. We will see in Chapter 4 that many solid-state ceramic processes involve simultaneous mass transport (diffusion), thermal transport, and reaction.

3.3

KINETIC PROCESSES IN POLYMERS

The previous sections dealt primarily with phase transformations and corrosion in materials. Polymers also undergo phase transformations. For example, there are many polymers that utilize nucleation and growth kinetics to transform from amorphous to crystalline polymers. The kinetics of these transformations are very similar, in principle, to the preceding descriptions for glasses, so it is not necessary to duplicate that material here. Polymers also are susceptible to corrosion, but the term degradation is more

KINETIC PROCESSES IN POLYMERS

247

widely utilized for deterioration in polymers. The kinetics of polymer degradation will be discussed in Section 3.3.3. Before we get to degradation, we should first discuss formation of polymers. Here we find kinetic processes that are unique to polymers. These are primarily formation reactions whose rates can vary widely depending on such variables as temperature and concentration. This is not the case with most metals. Thermodynamics tells us whether an alloy will form between two metals; and if that process is favorable, it generally proceeds spontaneously at a very rapid rate. The same is true of most ceramics and glasses: When the temperature is high enough, the reactants go to products—not only rapidly, but sometimes explosively. Polymer systems are more complex, in that the reacting molecules and products are much larger than those in metallic and ceramic systems, leading to mass- and heat-transport limitations that can affect the kinetics. It is now not enough to know that a reaction is thermodynamically favorable—we must look closely at the actual rate at which polymers are formed. This is the subject of Section 3.3.1. The effect of different types of processing condition on these reactions will then be discussed in Section 3.3.2.

3.3.1

Kinetics of Polymerization

Polymerization reactions are generally grouped into stepwise (also known as condensation) reactions and radical chain (sometimes called addition) reactions, depending upon the nature in which the monomers react with each other or with growing polymer chains to form higher-molecular-weight structures. Condensation reactions are characterized by the reaction of two functional groups that result in the loss of a small molecule. Thus, the repeat unit has a different molecular formula than the monomer from which it was formed. The small molecule formed during condensation is often water, though the specific formation of water is not a necessity for a condensation reaction. In contrast, addition polymerizations are those in which no loss of a small molecule occurs. This type of polymerization reaction usually requires an active center, such as a free radical or an ion, to proceed. Analysis of the kinetics for these two classes of reactions reveals some interesting differences not only in how the polymer is formed, but in how the reactions can be controlled. 3.3.1.1 Kinetics of Stepwise Polymerization. Consider the reaction of hexamethylene diamine and adipic acid: O H2N

(CH2)6

NH2 + HO

Hexamethylene diamine

O C

(CH2)4

C

OH

Adipic acid

kr

(3.60)

kf

O

O [HN

(CH2)6

H N

C

Nylon 66

(CH2)4

C]n

+ H2O

248

KINETIC PROCESSES IN MATERIALS

Note first of all that there is the characteristic loss of a small molecule for stepwise reactions, which in this case is water, and that the loss of water results in an amide (HN-CO) linkage; hence the resulting polymer is called a polyamide. This particular polyamide is very common and is better known by its trade name nylon. There are many types of nylons, depending on the exact formula of the diamine and diacid used to form the polyamide, but in this case, there are six carbon units between the nitrogens in the diamine and there are six carbon units (including the carbons which contain the double-bonded oxygens) in the diacid, so this polyamide is called Nylon 66. The kinetics of this reaction are analyzed by monitoring the disappearance of one of the reactants, or the formation of the product. If we choose to follow the disappearance of the adipic acid, the rate expression for a catalyzed reaction∗ is −d[COOH] = k[COOH][NH] dt

(3.61)

If we start with equal amounts of diamine and diacid, then for the duration of the reaction we have [COOH] = [NH] = c, and we have a second-order differential equation: −dc = kc2 dt

(3.62)

that is easily separated and integrated: 1 = kt + constant c

(3.63)

A convenient parameter for following the progress of these reactions is the fraction of functional groups reacted at any time, called the extent of reaction, p = (c0 − c)/c0 . Here, c0 is the initial concentration. In terms of the extent of reaction, the rate expression becomes 1 c0 kt = + constant (3.64) 1−p The quantity 1/(1 − p) is called the number-average degree of polymerization, x n , which represents the initial number of structural units present in the monomer relative to the total number of molecules, both monomer and polymer chains, present at any time t: x n = c0 kt + constant (3.65) Equation (3.65) shows us that the degree of polymerization increases linearly with time, an observation that has been verified experimentally. 3.3.1.2 Kinetics of Addition Polymerization. As the name suggests, addition polymerizations proceed by the addition of many monomer units to a single active center on the growing polymer chain. Though there are many types of active centers, and thus many types of addition polymerizations, such as anionic, cationic, and coordination polymerizations, the most common active center is a radical, usually formed at ∗

In the absence of an added strong acid, the reaction is self-catalyzed by a second molecule of the adipic −d[COOH] = k[COOH]2 [NH]. dt

acid, so the reaction is second order in acid:

KINETIC PROCESSES IN POLYMERS

249

Table 3.6 Characteristics of Stepwise and Addition Polymerization Reactions Stepwise Polymerization

Addition Polymerization

Any two molecular species present can react Monomer disappears early in the reaction Polymer molecular weight rises steadily throughout the reaction

Only growth reaction adds repeating units one at a time to the chain Monomer concentration decreases steadily throughout the reaction High polymer is formed immediately, polymer molecular weight changes little throughout the reaction

a double bond in the monomer, such as in vinyl chloride (see Table 3.6). As a result, addition polymerizations are sometimes called free radical polymerizations, or even vinyl polymerizations, although it should be noted that each of these terms is increasingly narrow in its definition. However, analysis of a vinyl polymerization provides a good example of the kinetics of addition polymerizations. In the 1930s, Paul Flory showed that radical polymerizations generally consist of three distinct steps: initiation, propagation, and termination. The free radical must first be initiated, then must propagate through the addition of monomers to the chain, and must eventually be terminated, either through consumption of the monomer or through addition of an agent that kills the free radicals. There are many ways to initiate radicals, such as ultraviolet radiation, electrochemical initiation, or oxidation–reduction reactions, but the most common is through addition of an initiator that transfers its radicals to the monomer to begin the polymerization. Organic peroxides, such as benzoyl peroxide, are common initiators that can transfer their radical to a monomer unit: O

O

O kd

f

C

C

O

O

C

C

C

2 f

f

C

ka H2C

O• CHCl

(3.66) O 2 f

C

C

H O

CH2

C• Cl

where φ is a phenol group. Most of these initiators have efficiencies of between 60% and nearly 100%, with recombination of the radical pairs the most common cause of low efficiency. If the fraction of radicals formed in the dissociation step that lead to successful radical chains is represented by f , then the rate of initiation is given by the change in initiated monomer concentration with time: d[Mž] = 2f kd [I] dt

(3.67)

where [I] is the concentration of the initiator, in this case benzoyl peroxide, and [Mž] is the radical concentration, in this case the initiated vinyl chloride monomer.

250

KINETIC PROCESSES IN MATERIALS

Once the monomer radical has been initiated, high polymer is formed through addition of monomer units to the radical in the propagation phase of the polymerization. Each time a monomer unit is added, the radical transfers to the end of the chain to allow the polymerization to continue: H R

(H2C

C•

CH2

CHCl)x

+ H2C

CHCl

Cl

kp

(3.68) H

R

(H2C

CHCl)x+1

CH2

C• Cl

The rate expression for propagation is given by the rate of monomer disappearance: −d[M] = kp [M][Mž] dt

(3.69)

where [M] is the instantaneous monomer concentration. It is assumed that the rate constant for propagation, kp , is the same for all propagation reactions in which a monomer is added to a growing radical chain, regardless of the molecular weight of the chain. The propagation reactions proceed at multiple radical sites until the monomer is exhausted, or until one of two common types of termination reactions occur: either combination or disproportionation. H CH2

H

H

C•

H

ktc

+



Cl

CH2

C

CH2

C

Cl

C

CH2

Cl

(3.70)

Cl

termination by combination H CH2

C• Cl

H +

H

H

ktd •

C

CH2

C

H + C

Cl

Cl termination by disproportionation

Cl

CH2

CH

(3.71)

Each of these termination reactions has a similar rate expression that is second order in radical concentration, so the overall termination rate can be expressed as the rate of disappearance of the radicals due to both types of termination: −d[Mž] = 2kt [Mž]2 dt

(3.72)

Other reactions can also take place, such as chain transfer reactions, in which the radical is transferred from a growing polymer chain to a small molecule such as the initiator

KINETIC PROCESSES IN POLYMERS

251

or a solvent molecule. These so-called chain transfer agents are often deliberately added to the reaction mix to either control the rate of the reaction or create branched polymers. The rate of the reaction can also be controlled with inhibitors or retarders. These additives also serve to help control the molecular weight of the polymer chain. In this instance, the rate of disappearance of monomer is given by an equation analogous to (3.69), except that the rate constant is now ktr,i ; and instead of adding monomer to continue chain propagation, Mž now transfers its radical to some agent, A, which may be an initiator, a solvent, a monomer, or a chain-transfer agent: −d[M]  = ktr,i [A][Mž] dt

(3.73)

In the steady-state condition, the rate of initiation is equivalent to the rate of radical termination, and the radical concentration, [Mž], remains essentially constant with time. Reactions (3.67) and (3.72) can be set equal at steady state to solve for [Mž], [Mž] =



f kd [I] kt

1/2

(3.74)

which is then substituted back into Eq. (3.69) to obtain the overall rate expression for addition polymerization: −[M] = kp dt



f kd [I] kt

1/2

[M]

(3.75)

As was the case for condensation-type reactions, there is a number-average degree of polymerization, x n , where xn =

rate of growth rates of reactions leading to dead polymer

(3.76)

The rate of growth is simply given by Eq. (3.75), and the sum of the reaction rates leading to dead polymer come from termination [Eq. (3.72)] and chain transfer [Eq. (3.73)] reactions kp [Mž][M] xn = (3.77)  2 ktr,i [A][Mž] ϒkt [Mž] + i

where ϒ = 1 for termination by coupling and 2 if termination is by disproportionation, since two dead polymer molecules are formed in the latter reaction but only one dead polymer molecule is formed in coupling. Equation (3.75) is more easily simplified by inverting it to obtain an expression for 1/x n :

1 = xn

ϒkt [Mž]2 +



ktr,i [A][Mž]

i

kp [Mž][M]

=

  ϒkt [Mž]  ktr,i [A] + kp [M] kp [M] i

(3.78)

252

KINETIC PROCESSES IN MATERIALS

Cooperative Learning Exercise 3.4 Work in groups of three. Equation 3.75 can be re-written to reflect the fact experimental determination of polymerization rates typically yield a single rate constant, called the apparent rate constant, kapp , which is a composite of the three different rate constants, kd , kt and kp . If we call the rate of disappearance of monomer concentration, −d[M]/dt the rate of polymerization, Rp , then we can rewrite Equation 3.75 as

Rp = kapp [M][I]1/2 or ln Rp = ln kapp + ln[M] + 0.5 ln[I] This relationship will allow us to estimate the variation in polymerization rate with temperature. Taking the derivative and treating the monomer and initiator concentrations as constants with respect to temperature gives d ln Rp = d ln kapp or (dRp )/Rp = d ln kapp From here, we can introduce temperature dependence by assuming that kapp is given by an Arrhenius-type expression (Equation 3.12). 1. Work together to derive the following expression for the temperature-dependence of the polymerization rate from the previous expression: Ea,t Ea,d Ea,p + + dRp 2 2 dT = Rp RT 2 where Ea,p , Ea,d , and Ea,t are the activation energies for propagation, initiator decomposition, and termination, respectively. 2. Use the data below for the activation energies of some typical addition polymerization reactions at 60◦ C to calculate an average activation energy for propagation, termination, and initiator decomposition. Each person should calculate one of the three average activation energies.

Monomer Acrylonitrile Methyl acrylate Methyl methacrylate Styrene Vinyl acetate 2-Vinyl pyridine

Ea,t Ea,p (kJ/mol) (kJ/mol) 16.2 29.7 26.4 26.0 18.0 33.0

15.5 22.2 11.9 8.0 21.9 21.0

Initiator 2,2′ Azobisisobutyronitrile in benzene t-Butyl peroxide in benzene Benzoyl peroxide in benzene t-Butyl hydroperoxide in benzene

Ea,d T(◦ C) (kJ/mol) 70

123.4

100

146.9

70

123.8

169

170.7

3. Combined your answers to estimate the percent rate of change in the polymerization rate (dR p /Rp × 100) for each 1◦ C change in temperature (dT = 1) at 50◦ C. Answer : Ea,p = 24.9; Ea,t = 16.8; Ea,d = 141; dRp /Rp = {[24.9+141/2+16.8/2](103 )(1)/(8.314)(323)2 } × 100 = 12%◦ C−1

KINETIC PROCESSES IN POLYMERS

253

Once again, elimination of [Mž] using Eq. (3.74) gives a general expression for the instantaneous degree of polymerization as a function of rate constants and concentrations:   1/2 ϒf kd kt [I]  ktr,i [A] 1 + = (3.79) xn kp [M] kp [M] i The first term in Eq. (3.79) is the inverse of the degree of polymerization that would occur in the absence of chain transfer, (x n )0 . The second term represents the effect of different chain transfer agents on the molecular weight. 3.3.1.3 Kinetics in the Polymerization of Copolymers. Recall from Chapter 1 that a copolymer is a polymer composed of two or more different repeat units—the result of the simultaneous polymerization of two or more monomers. Though copolymers can be formed by any of the reaction mechanisms described above, the most interesting copolymerizations are of the free radical type. This is because the reactivity of the radicals can vary depending on the structure of the monomer to which they are attached. Such changes in reactivity between monomers are not generally found in condensation-type reactions; that is, a carboxylic acid group is a carboxylic acid group, independent of the chain or monomer to which it is attached. In chain polymerizations, however, the monomers can add to the growing chain in different ways and with different rates. Let us examine the simple system of two monomers M1 and M2 that are initiated to form radicals M1 ž and M2 ž. There are four ways in which a monomer can add to a growing chain and four corresponding rate expressions: k11

M1 ž + M1 −−−→ M1 . k12

M1 ž + M2 −−−→ M2 . k21

M2 ž + M1 −−−→ M1 . k22

M2 ž + M2 −−−→ M2 .

(3.80a) (3.80b) (3.80c) (3.80d)

The rate of disappearance of each monomer is then given by −d[M1 ] = k11 [M1 ž][M1 ] + k21 [M2 ž][M1 ] dt −d[M2 ] = k12 [M1 ž][M2 ] + k22 [M2 ž][M2 ] dt

(3.81) (3.82)

If we define the monomer reactivity ratio for monomer 1 and 2, r1 and r2 , respectively, as the ratio of rate constants for a given radical adding to its own monomer to the rate constant for it adding to the other monomer (r1 = k11 /k12 and r2 = k22 /k21 see Table 3.7 for typical values), then we arrive at the following relationship known as the copolymer equation: [M1 ](r1 [M1 ] + [M2 ]) d[M1 ] = (3.83) d[M2 ] [M2 ]([M1 ] + r2 [M2 ])

254

KINETIC PROCESSES IN MATERIALS

Table 3.7 Typical Monomer Reactivity Ratios in the Addition Polymerization of Copolymers Monomer 1

Monomer 2

r1

r2

T(◦ C)

Acrylonitrile

1,3-Butadiene Methyl methacrylate Styrene Vinyl acetate Vinyl chloride Methyl methacrylate Styrene Vinyl chloride Styrene Vinyl acetate Vinyl chloride Vinyl acetate Vinyl chloride Vinyl chloride

0.02 0.15 0.04 4.2 2.7 0.75 1.35 8.8 0.46 20 10 55 17 0.23

0.3 1.22 0.4 0.05 0.04 0.25 0.58 0.035 0.52 0.015 0.1 0.01 0.02 1.68

40 80 60 50 60 90 50 50 60 60 68 60 60 60

1,3-Butadiene

Methyl methacrylate

Styrene Vinyl acetate a

Data from Lewis J. Young, Copolymerization Reactivity Ratios, pp. II-105–II-386 in J. Brandrup and E. H. Immergut, eds., with the collaboration of W. McDowell, Polymer Handbook, 2nd ed., WileyInterscience, New York, 1975.

Though this is a useful relationship, there are some special cases that are more easily described by introducing two types of ratios. The first ratio is F1 , the rate of change of monomer M1 mole fraction: F1 =

d[M1 ]/dt d[M1 ]/dt + d[M2 ]/dt

(3.84)

F2 =

d[M2 ]/dt d[M1 ]/dt + d[M2 ]/dt

(3.85)

Similarly,

The second useful ratio is the instantaneous concentration of a given monomer, f1 or f2 : [M2 ] [M1 ] f1 = , f2 = (3.86) [M1 ] + [M2 ] [M1 ] + [M2 ] It is important to distinguish between the concentration ratio, fi , and the rate of change of concentration, Fi , since if monomers M1 and M2 are consumed at different rates, then Fi = fi . Substitution of these ratios F1 , F2 , f1 , and f2 into the copolymer equation gives (r1 f1 /f2 ) + 1 F1 = (3.87) F2 (r2 f2 /f1 ) + 1 or

[r1 f1 /(1 − f1 )] + 1 F1 = 1 − F1 [r2 (1 − f1 )/f1 ] + 1

(3.88)

KINETIC PROCESSES IN POLYMERS

255

This form of the copolymer equation allows us to identify several simplifying cases: ž

k12 and k21 are negligible; that is, a radical prefers to react with its own monomer, and no copolymer is formed—only homopolymer. ·B-B-B-B-B A-A-A-A-A·

ž

k11 and k22 are negligible; that is, a radical prefers to react with the other monomer, F1 = F2 = 0.5, ·B-A-B-A-B

A-B-A-B-A·

ž

and a perfectly alternating copolymer is formed. k11 ≈ k12 and k22 ≈ k21 ; r1 = r2 = 1; F1 = f1 , and a perfectly random copolymer is formed. ·B-A-B-A-A A-B-B-B-A·

ž

r1 r2 = 1, so that Eq. (3.88) becomes r1 f1 F1 = 1 − F1 1 − f1

3.3.2

(3.89)

Polymerization Processes∗

For the engineer, a knowledge of the reaction kinetics is merely a piece of what is usually a much larger problem involving the design of reactor components and the production and processing of large quantities of product—in this case, polymers. As a result, it is useful to discuss not only how the kinetic principles developed so far are applied to different methods of polymerization, but how problems associated with real processes, such as heat generation and removal, can in turn affect the polymerization kinetics (see Table 3.8). We will look at four common polymerization processes: bulk polymerization, solution polymerization, suspension polymerization, and emulsion polymerization. Though most of these processes are used primarily for addition-type polymerizations, applications involving condensation-type polymerizations will be noted where appropriate. 3.3.2.1 Bulk Polymerization. Monomer and polymer (with traces of initiator) are the only constituents in bulk polymerizations. Obviously, the monomer must be soluble in the polymer for this type of process to effectively proceed. Bulk polymerization, also called mass or block polymerization, can occur in stirred-tank reactors, or can be unstirred, in which instance it is called quiescent bulk polymerization. The primary difficulty with bulk polymerizations is that as the polymerization proceeds and more polymer is formed, the viscosity increases, thermal conductivity decreases, and heat removal becomes difficult.

256

KINETIC PROCESSES IN MATERIALS

Table 3.8

Comparison of Polymerization Processes

Type Batch bulk Continuous bulk Solution

Advantages

Disadvantages

Minimum contamination

Poor heat control

Simple equipment

Broad MW distribution

Better heat control Narrower MW distribution Good heat control Direct use of solution possible

Requires stirring, separation, and recycle Not useful for dry polymer Solvent removal difficult

Suspension

Excellent heat control Direct use of suspension possible

Requires stirring Contamination by stabilizer possible Additional processing (washing, drying) required

Emulsion

Excellent heat control

Contamination by emulsifier likely

Narrow MW distribution High MW attainable

Additional processing (washing, drying) required

Direct use of emulsion possible Source: Adapted from F. W. Billmeyer, Textbook of Polymer Science, Wiley, New York.

There is an increase in temperature as heat removal slows, along with a corresponding increase in the reaction rate. This phenomenon is known as the autoacceleration or Trommsdorff effect and can lead to catastrophic results if not properly controlled. Even with successful control of the reaction, it is difficult to remove the traces of remaining monomer from the polymer due to decreased diffusion. Similarly, it is difficult to get the reactions to proceed to completion due to limited monomer mobility. A bulk polymerization reactor can be as simple as a tube into which the reactants are fed and from which the polymer mixture emerges at the end; it can be more of a traditional, continuous stirred-tank reactor (CSTR), or even a high-pressure autoclave-type reactor (see Figure 3.21). A bulk polymerization process need not be continuous, but it should not be confused with a batch reaction. There can be batch bulk polymerizations just as there are continuous bulk polymerizations processes. 3.3.2.2 Solution Polymerization. By adding a solvent to the monomer–polymer mixture, heat removal can be improved dramatically over bulk reactions. The solvent must be removed after the polymerization is completed, however, which leads to a primary disadvantage of solution polymerization. Another problem associated with radical chain polymerizations carried out in solution is associated with chain transfer to the solvent. As we saw in Section 3.3.1.2, chain transfer can significantly affect the molecular weight of the final polymer. This is particularly true in solution polymerization, where there are many solvent molecules present. In fact, chain transfer to solvent often dominates over chain transfer to other types of molecules, so that Eq. (3.79) reduces to

1 = xn



1 xn



0

+

CS [S] [M]

(3.90)

KINETIC PROCESSES IN POLYMERS

257

Fresh monomer

Vent

Nitrogen Prepolymerizer

Vacuum Compressor Vent

Dust resin separator

Recycle vinyl chloride condenser

Vinyl chloride from storage

Autoclave Monomer feed tank PVC resin to grinding and bagging

Figure 3.21 Schematic diagram of vinyl chloride bulk, continuous polymerization. Reprinted, by permission, from A. Krause, Chem. Eng., 72, p. 72. Copyright  1965 by McGraw-Hill.

where the first term once again describes the contributions of propagation to the molecular weight, [S] and [M] are concentrations of the solvent and monomer, respectively, and CS is the ratio of rate constants ktr,s /kp . This equation shows that a plot of 1/x n versus [S]/[M] should yield a straight line, the intercept for which gives the degree of polymerization in the absence of chain transfer to solvent. This relationship has been borne out experimentally for polystyrene and several solvent systems (see Figure 3.22). The reactor equipment used for solution polymerizations is typically glass-lined stainless steel. An example of solution polymerization is the reaction of ethylene in isooctane with a chromia silica alumina catalyst initiator (see Figure 3.23) to form polyethylene. Typical reaction conditions for this polymerization are 150–180◦ C and 2.1–4.8 MPa (300–700 psi). 3.3.2.3 Suspension Polymerization. When water is the solvent and the monomer is insoluble in water, suspension polymerization is often employed. The monomer and initiator form small droplets within the aqueous phase, and polymerization proceeds via bulk polymerization. Droplet sizes are typically between 10 and 1000 µm

258

KINETIC PROCESSES IN MATERIALS

200

150

105 100 xn

50

Transfer agent Ethylbenzene Isopropylbenzene Toluene Benzene

0

0

10

20

30

[S]/[M]

Figure 3.22 Effect of chain transfer to solvent on number average degree of polymerization for polystyrene [data from R. A. Gregg and F. R. Mayo, Discuss Faraday Soc., 2, 328 (1947).] From F. W. Billmeyer, Textbook of Polymer Science, 3rd ed. Copyright  1984 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Isooctane

CH2 = CH2

Catalyst bed

Separator Recycle

Polymer

Figure 3.23

Schematic diagram of ethylene solution polymerization process.

259

KINETIC PROCESSES IN POLYMERS

in diameter, so that heat transfer to the suspending medium and subsequently to the reactor walls is very efficient throughout the course of the reaction. The result of the polymerization within each of the droplets is a polymer “bead.” Though some agglomeration of particles takes place, the addition of a suspending agent such as poly(vinyl alcohol) minimizes coalescence, and the final polymer bead diameters are typically 100–1000 µm. The polymer beads are recovered by filtration, which is followed by a washing step. Suspension polymerization can be carried to nearly 90% completion at moderate pressures of 600 kPa (88 psi) and 80◦ C with stirring at about 150 to 300 rpm. A schematic diagram of the suspension polymerization of polystyrene is shown in Figure 3.24, and one for methyl methacrylate is shown in Figure 3.25. 3.3.2.4 Emulsion Polymerization. Emulsion polymerization is similar to suspension polymerization in that the monomer is insoluble in the solvent, water. The similarity ends there. In emulsion polymerization, the initiator is water-soluble, the particles in which the propagation step occurs are smaller, typically 0.05–5 µm in diameter, and the final beads are not usually filtered from the aqueous solution. The emulsifier, or surfactant, plays a very important role in emulsion polymerization. It is usually a long-chained hydrocarbon with a hydrophylic “head” and hydrophobic “tail.” This charge polarization causes the tails to surround the monomer molecules, with the heads pointing outward into the water phase, forming a spherical droplet called a micelle (see Figure 3.26). As radicals initiate in the aqueous phase, they react with what few monomer molecules are dissolved in the water. As the radicals slowly grow and become more hydrophobic, they find it more stable to be inside of the micelles. Upon entering a micelle, the radicals cause rapid propagation and the micelle expands. It can be shown that the degree of polymerization in emulsion polymerization can be derived from general polymerization kinetics [Eq. (3.79)] by neglecting chain transfer:

2f kd [I] 1 = xn kp [M∗ ][N∗ ] Suspension agents Styrene Water

Water-wash tank

Dried polystyrene beads

Continuous solid-bowl centrifugal Waterpolystyrene slurry Water Batch polymerization reactor (4)

(3.91)

Storage and blending tank (4)

Polystyrene to pelletizing and storage Hopper

Rotary dryer Polystyrene extruder

Cooler

Figure 3.24 Schematic diagram of styrene suspension polymerization process. Reprinted, by permission, from Chem. Eng., 65, 98. Copyright  1958 by McGraw-Hill.

260 Discharge conveyor

Washing centrifuge

Cooling water

Cond.

Color mix section

Reject material

Quality control and blending section

Beads

Scrap Discharge screen

Steam

Fluidized bed dryer Air

Wet air

Hopper

Extruder feed bins

Vent

Air conveyer

Water through

Molding powder

Scale

Vibrating screen Drumming hopper

Feed hopper Pelletizer Vacuum pump Strand die Air knife Extruder

Airjet

Colored beads

Clear beads

Figure 3.25 Schematic diagram of methyl methacrylate suspension polymerization process. Reprinted, by permission, from E. Guccione, Chem. Eng., 73, 138. Copyright  1966 by McGraw-Hill.

Wash water

Centrifuge feed tank

Slurry tanks

Water Chemicals

Nitrogen Water mix tank

Steam Cooling water Polymerization Reactors

Monomer mix tank

Methyl methacrylate

Benzoyl peroxide

KINETIC PROCESSES IN POLYMERS

261

Micelle with solubilized monomer

Monomer droplet

Emulsifier in aqueous solution

Swollen polymer particle

Figure 3.26 Schematic representation of micelle formation in emulsion polymerization.

where the superscript “∗ ” on the monomer concentration indicates that this is the concentration of monomer in the micelle, not the solution. Since the initiator is soluble in water, its concentration remains based upon the total volume of water present. In Eq. (3.91), [N∗ ] is the concentration of micelle-swollen polymer particles. The final product of emulsion polymerization is an emulsion —a stable, heterogeneous mixture of fine polymer beads in an aqueous solution, sometime called a latex emulsion. Water-based paints, for example, can be formed from the emulsion polymerization of vinyl acetate. In this process, 1 m3 of water containing 3% poly(vinyl alcohol) and 1% surfactant are heated to 60◦ C in a reaction vessel (see Figure 3.27) The temperature rises to around 80◦ C over a 4 to 5 hour period as monomer and an aqueous persulfate solution are added. The rate at which heat can be removed limits the rate at which monomer can be added. Copolymerizations can also be carried out in an emulsion reaction. For example, nitrile rubber (NBR) can be made from the emulsion polymerization of acrylonitrile and butadiene: H2C

C

CN + H2C

Acrylonitrile

CH

HC

CH2

Butadiene

(3.92) [

H2C

CH

CH2

CH

CH

CH2

]n

CN

As the acrylonitrile content in nitrile rubber increases, so does the resistance to nonpolar solvents. 3.3.2.5 Gas-Phase Polymerization. Not all polymerization reactions are carried out in the liquid phase. Polyethylene, for example, can be polymerized not only in solution as discussed in Section 3.3.2.2, but also through gas-phase polymerization (see Figure 3.28). Significant economic advantage is gained by eliminating costly solvent and catalyst recovery equipment. A fluidized bed reactor is used, into which purified

262

KINETIC PROCESSES IN MATERIALS

Monomer Water Vacuum inlets inlet

Inlet manifold Pressure gauge

Stirrer motor Oil seal Oil in Oil out

To exhaust hood Manhole

Steam Bursting disk Valves operated by temperature controller

Offset stirrer Heating jacket Water Outlet

Thermocouple to temperature controller

Main outlet

Sampling outlet

Figure 3.27 Schematic diagram of a typical emulsion polymerization reactor. Reprinted, by permission, from J. A. Brydson, Plastic Materials, p. 160. Copyright  1966 by Van Nostrand.

Cycle compressor Compressor Cycle cooler Fluid-bed reactor Purge Catalyst feeder

Ethylene Comonomer

Product discharge system

Nitrogen Granular polyethylene to blending and storage

Figure 3.28 Schematic diagram of ethylene gas phase polymerization process, Reprinted, by permission, from Chem. Eng., 80, 71. Copyright  1979 by McGraw-Hill.

KINETIC PROCESSES IN POLYMERS

263

ethylene and a modified chromia on silica powdered catalyst are fed. The reactor pressure is about 2 MPa (290 psi) and the temperature is 85–100◦ C. Heat is removed as gas is circulated using external coolers. The growing particles remain in the reactor 3–5 hours; emerging with an average diameter of about 500 µm. Small amounts of catalyst are used, so catalyst removal is often not necessary. Hydrogen is used as a chain transfer agent, and branching can be introduced by the addition of 1-butene. 3.3.3

Kinetics of Polymer Degradation

Degradation of polymers may be considered as any type of undesired modification of a polymer chain backbone, side groups, or both. These modifications are usually chemical in nature; that is, they involve the breaking of primary bonds, leading to reduced molecular weight and altered physical properties. The causes of degradation may be environmental (such as heat, light, or atmospheric pollutants), induced, (as in rubber mastication), or a result of processing due to heat and/or air. We will be concerned primarily with environmental degradation. Even within environmental degradation, there are six primary classes of polymer degradation: thermal, oxidative, radiative, chemical, mechanochemical, and biological. We will look briefly at the mechanisms of each of these types of degradative processes and present without derivation their related kinetic expressions. 3.3.3.1 Thermal Degradation. In thermal degradation, a polymer chain—in this case, a vinyl polymer chain—depolymerizes, or unzips. Just as in the formation reactions, there is initiation, chain transfer, and termination. Once a free radical has been initiated through some sort of random or chain-end scission process, the depropagation step proceeds, and monomer fragments split off from the polymer chain in a stepwise manner (see Eq. 3.93). The monomer fragments are usually assumed to be volatile. H R

(H2C

CHCl)x+l

CH2

C•

H kdp

R

(H2C

CHCl)x

CH2

C•

(3.93) Cl

Cl

+ H2C

CHCl

It can be shown that the change in polymer weight, W , with time, t, for thermal degradation is given by −dW 2kir b(x n )0 W (2b−1)/b = (b−1)/b dt W0

(3.94)

where kir is the rate constant for random initiation of radicals, W0 is the initial weight of the polymer, (x n )0 is the initial degree of polymerization, and b is a parameter that depends on the degree of polydispersity (b = 1 for a monodispersed system). The value of b is a measure of the breadth of the molecular weight distribution curve. For example, when the molecular weight distribution follows the exponential distribution, b = 2. According to Eq. (3.94), this corresponds to a reaction order of 3/2. As b increases, the reaction order approaches a limiting value of 2.

264

KINETIC PROCESSES IN MATERIALS

3.3.3.2 Oxidative Degradation. There is a wealth of literature available on the oxidatative degradation of polymers. Unfortunately, there is no single mathematical treatment that is applicable to all polymer types. As a result, the development of an appropriate kinetic expression is highly dependent upon the type of polymer being considered and the presence of additives, inhibitors, and antioxidants. For the sake of simplicity, we will consider a general case of autoxidation in a polyolefin in the absence of additives. It will be assumed that the diffusion of oxygen is rapid and is therefore not the rate-limiting kinetic step. As in thermal degradation, the process begins with an initiation step, in this case the formation of radicals where X denotes a source of free radicals, such as polymer chains, which then react with atmospheric oxygen: k1

X −−−→ R ž k2

Rž + O2 −−−→ RO2 ž

(3.95a) (3.95b)

The oxidized radicals can then react with another polymer chain, RH, to form peroxides: k3

RO2 ž + RH −−−→ RO2 H + Rž

(3.96)

Reactions (3.95) and (3.96) are the propagation steps, although the former is very rapid, with a negligible activation energy, whereas the latter is slower, with an activation energy of about 1.7 kJ/mol (7 kcal/mol). Three types of termination reactions ensue, all of which are very rapid, with activation energies of less than 1 kJ/mol (3 kcal/mol): k4

2Rž −−−→ Products k5

RO2 ž + Rž −−−→ Products k6

2RO2 ž −−−→ Products + O2

(3.97a) (3.97b) (3.97c)

In the presence of oxygen, the first two termination reactions can be neglected. The oxidation rate can then be calculated from the rate-limiting propagation step, Eq. (3.96): d[RO2 H] = k3 [RO2 ž][RH] dt

(3.98)

A steady-state approximation allows us to solve for the intermediate free-radical concentration, [RO2 ž], to obtain 1/2

k3 k1 [X]1/2 [RH] d[RO2 H] = 1/2 dt k6

(3.99)

3.3.3.3 Radiative Degradation. Radiative degradation in polymers can give rise to both low-molecular-weight species, as in thermal degradation, or crosslinked structures, which are insoluble and infusible. Though crosslinking actually leads to an

KINETIC PROCESSES IN POLYMERS

265

increase in the molecular weight of the sample, it is still considered a degradative process since the resulting product has unfavorably altered physical properties. There are generally two types of radiative degradation processes in polymers, grouped according to the amount of radiant energy involved and which result in different types of degra˚ imparts dation. Photolysis occurs when ultraviolet light (wavelength λ = 102 –104 A) energy on the sample of the order 102 –103 kJ/mol. Radiolysis occurs when higher ˚ impart energy of the order 105 –1010 kJ/mol to the radiant waves (λ = 10−3 –102 A) sample. Each of these types attacks the polymer chain in a different fashion. 3.3.3.4 Chemical Degradation. For most applications in the chemical process industry, chemical degradation is the most important type of polymer degradation. It is also the largest class overall, encompassing degradation due to both gaseous and liquid species on all polymer classes. Even within a polymer class, such as polyolefins, slight changes in the chain chemistry can lead to enormous changes in chemical compatibility. As a result, the presentation here must be severely limited. The intents will be to provide some insight into how the chemical degradation occurs, and to present practical information in guiding the selection of an appropriate polymer for a specific chemical environment. In addition to oxygen, the most important gaseous agents that lead to polymer chemical degradation are nitrogen dioxide (NO2 ), sulfur dioxide (SO2 ), and ozone (O3 ). Nitrogen dioxide attacks hydrocarbons by removing a tertiary hydrogen atom:

NO2 (g) + RH

HNO2 + Rž

(3.100)

The resulting radical then reacts further with nitrogen dioxide, NO2 (g) + Rž

RNO2

(3.101)

which then leads to products via chain scission and/or crosslinking. Polystyrene undergoes such a degradation process, the degree of polymerization for which changes according to the following empirical relationship 1 1 − = kt xn (x n )0

(3.102)

where the rate constant, k, follows an Arrhenius-type relationship of the form k = PNO2 [41.4 exp(−3870/RT )]hr−1

(3.103)

Here PNO2 is the partial pressure of nitrogen dioxide in pascals (1 Pa = 0.000145 psi), and the activation energy is 3870 kJ/mol. Other saturated polymers are less susceptible to attack by NO2 than most unsaturated polymers such as synthetic rubbers (polyisoprene, polybutadiene, and butyl rubber). The presence of oxygen also tends to accelerate degradation by NO2 . The reaction of sulfur dioxide with saturated polymers is complex, but appears to be activated by ultraviolet radiation. 3.3.3.5 Mechanochemical Degradation. Mechanochemical degradation occurs in polymers as the result of an applied mechanical force. This type of degradation is quite common in machining processes such as grinding, ball milling, and mastication.

266

KINETIC PROCESSES IN MATERIALS

In general, the degradation processes are similar to the other types of degradation, in that radicals are formed and cause subsequent reactions, such as oxidation and chain transfer, that can lead to either crosslinking or a reduction in molecular weight, though molecular weight reduction is the most common result. There are some unique aspects of mechanochemical degradation, however. In contrast to thermal degradation, radical formation is nonrandom, and the location of chain scission depends on the manner in which the forces are applied to the polymer. Also, the molecular weight decreases rapidly from an initial degree of polymerization, (x n )0 , for polymers under mechanical shear, then level off to a limiting degree of polymerization, (x n )∞ (Figure 3.29). This exponential decrease in molecular weight and degree of polymerization with time can be expressed by an empirical relationship of the form (x n )0 − (x n )t = 1 − exp(−kt) (x n )0 − (x n )∞

(3.104)

where (x n )t is the degree of polymerization at time t. Typical values of k for common polymers of various molecular weights are given in Table 3.9. In addition to a reduction in molecular weight, mechanical forces generally lead to an increase in solubility, a narrowing of the molecular weight distribution, decreased tensile strength, decreased crystallinity, and increased plasticity. These effects are highly dependent upon the machining conditions. In particular, higher machining temperatures result in non-Arrhenius behavior since there is a corresponding decrease in viscosity. The presence of radical scavengers and the type of machining equipment employed

M × 10−3

120

80

40

M∞ 0

20

40 60 Time, hours

80

Figure 3.29 Effect of machining time on molecular weight of polymer.

Table 3.9 Rate Constants for Mechanochemical Chemical Degradation of Polymers Polymer Poly(vinyl alcohol) Polystyrene Poly(methyl methacrylate) Poly(vinyl acetate)

Molecular Weight 4,000 7,000 9,000 11,000

k, hr−1 0.0237 0.0945 0.1200 0.0468

KINETIC PROCESSES IN POLYMERS

267

also play significant roles in determining the extent of mechanochemical degradation during machining. 3.3.3.6 Biological Degradation. In contrast to the degradation processes discussed to this point, biological degradation in polymers is oftentimes advantageous. In fact, there is a class of polymers that have been developed to enhance their susceptibility to attack by biological agents in order to improve their biodegradability and reduce their long-term impact on the environment. Other important applications of biodegradable polymers are as biomaterials, such as resorbable sutures, and for controlled drug delivery devices. There are generally two types of polymeric materials that are susceptible to biological agents such as water, enzymes, and microbes: (a) natural materials like collagen, cellulose, and starch and (b) synthetic polymers. Though natural polymers represent a large fraction of all biodegradable polymers, particularly for biomedical uses like catgut sutures, we will concentrate on synthetic polymers, since they represent the class of materials where most of the advancement is still possible. It is the synthetic polymers that can be optimized in terms of mechanical properties and rate of degradation for biological and biomedical applications. Substitutes for petroleum based polymers used in packaging (e.g., polystyrene, polyethylene, and polypropylene) represent one of the largest potential markets for biodegradable polymers. Some natural polymers have been used as substitutes, but their processability and mechanical properties have limited their use. In addition, factors such as cost, recycling, and alternative strategies like incineration have limited the need for the development of synthetic biodegradable polymers. Work continues in these areas, but it is the realm of synthetic polymers for biomedical applications that represents one of the more important areas of new material development. It is also the area where the kinetics of the degradation processes have been most widely studied. Though several classes of biodegradable synthetic polymers have been developed, including polyamides based on 4, 4′ -spirobibutyrolactone [13] and poly(hydroxylalkanoates), it is the poly(α-esters) that have been most widely utilized for biomedical applications. Let us look at a specific example of synthetic polymers for resorbable suture material. The most common synthetic biodegradable polymers for suture material and their corresponding weight loss in aqueous solution are listed in Table 3.10. Of these, poly(glycolic acid), PGA, poly(lactic acid), PLA, and copolymers of these two polyesters are the most widely used for resorbable suture material. PGA is a tough, Table 3.10 Comparative Weight Loss for Some Common Biodegradable Polymers, Given as Time for 10% Weight Loss at 37◦ C and pH 7.4 Polymer Poly(L-lactic acid) (PLLA 8%)-co-Poly(glycolic acid) (PGA 92%) Poly(glycolic acid) (PGA) Poly(p-dioxanone) (PDS) Poly(3-hydroxybutyrate-co-3-valerate) PHBV (20% hydroxyvalerate, HV) PHBV (12% HV) Poly(3-hydroxybuturate) HB (O% HV) Source: W. Amass, A. Amass, and B. Tighe, Polym. Int., 47, 89–144 (1998).

t10 (hr) 450 550 1200 4.7% @ 5500 5.6% @ 5500 1.8% @ 2500

268

KINETIC PROCESSES IN MATERIALS

crystalline, inelastic material with a Tm of 230◦ C and Tg of 36◦ C. There are two types of poly(lactic acid): poly(L-lactic acid), PLLA, which is a tough, crystalline, inelastic material with Tm = 180◦ C, Tg = 67◦ C; and poly(D-lactic acid), PDLA, which is a tough, amorphous, inelastic material with Tg = 57◦ C. The in vitro and molecular weight loss data for PLLA show that over a 16-week period, only 10–15% of the mass is lost, and the molecular weight is reduced to ∼50% of its original value [14]. Similar studies for PGA show that mass loss begins in the fourth week and is completed in 10–12 weeks (see Figure 3.30 for two different types of commercial PGA sutures), and the molecular weight loss data (Figure 3.31) shows a steady decrease in the molecular weight of the primary peak. The degradation mechanisms of these polymers are quite complex. Factors influencing the biological stability include the molecular weight (high MW gives stability), the presence of residual monomer and low-molecular-weight compounds, temperature, % crystallinity (high crystallinity decreases the rate of biodegradation), and extent of sterilization by β- and γ -radiation (both lead to degradation). Polyesters generally undergo a

100

(a)

80

% mass remaining

60 40 20 0 100

(b)

80 60 40 20 0

1

2

3 4 5 Time (weeks)

6

7

8

Figure 3.30 Percent weight loss for two types of PGA sutures, (a) Dexon and (b) Dexon-s, as function of exposure time to a buffered solution (pH 7, 37◦ C). Reprinted, by permission, from J. Feijin, in Polymeric Biomaterials, p. 68. Copyright  1986 by Martinus Nijhoff Publishers.

14 days

56 days 7 days 21 days

0 days

28 days

Low MW

High MW

Figure 3.31 Shift in molecular weight distribution for a commercial PGA suture material (Dexon) as a function of exposure time to a buffered solution (pH 7, 37◦ C). Reprinted, by permission, from J. Feijin, in Polymeric Biomaterials, p. 68. Copyright  1986 by Martinus Nijhoff Publishers.

KINETIC PROCESSES IN COMPOSITES

269

two-stage hydrolysis process. In the first stage, water diffuses into the amorphous regions, producing random hydrolytic scission at the susceptible ester linkage. This leads to an apparent increase in the percent crystallinity of the polymer during this stage. After most of the amorphous regions have been eroded, hydrolysis of the crystalline regions commences. A method for determining the mode of hydrolysis has been developed by Shih [15], in which the mole fraction of monomer, m1 , and the degree of polymerization, α, are determined by a technique like nuclear magnetic resonance (NMR) spectroscopy. If completely random scission of the backbone bonds occurs, m1 = α 2 , and if exclusive chain-end unzipping occurs, m1 = α. In addition to hydrolysis, these polymers can undergo degradation due to the effect of enzymes, chemicals, and radiation.

3.4

KINETIC PROCESSES IN COMPOSITES∗

Due to the fact that industrial composites are made up of combinations of metals, polymers, and ceramics, the kinetic processes involved in the formation, transformation, and degradation of composites are often the same as those of the individual components. Most of the processes we have described to this point have involved condensed phases—liquids or solids—but there are two gas-phase processes, widely utilized for composite formation, that require some individualized attention. Chemical vapor deposition (CVD) and chemical vapor infiltration (CVI) involve the reaction of gas phase species with a solid substrate to form a heterogeneous, solid-phase composite. Because this discussion must necessarily involve some of the concepts of transport phenomena, namely diffusion, you may wish to refresh your memory from your transport course, or refer to the specific topics in Chapter 4 as they come up in the course of this description. 3.4.1

Kinetics of Chemical Vapor Deposition∗

Chemical vapor deposition (CVD) involves the reaction of gaseous reactants to form solid products. There are two general types of CVD processes: (a) the thermal decomposition of a homogeneous gas to form a solid and (b) the chemical reaction of two or more gaseous species to form a solid. Both types of CVD reactions are used industrially to form a variety of important elements and compounds for semiconductor, superconductor, and ceramic coating applications. An example of the first kind can be found in the decomposition of trichloromethylsilane to give silicon carbide and hydrogen chloride: (CH3 )SiCl3 (g)

SiC(s) + 3HCl(g)

(3.105)

The reaction of titanium tetrachloride with boron trichloride and hydrogen to give solid titanium diboride is an example of the second type of CVD reaction: TiCl4 (g) + 2BCl3 (g) + 5H2 (g)

TiB2 (s) + 10HCl(g)

(3.106)

Reaction (3.106) illustrates the primary advantage of a CVD process. Solid titanium diboride, TiB2 , melts at 3325◦ C, yet it can be produced via reaction (3.106) at 1027◦ C or lower on a suitable substrate by the CVD process. Additionally, the diffusion of gas

270

KINETIC PROCESSES IN MATERIALS

phase reactants is generally rapid, allowing for the efficient coating of high-surfacearea (i.e., porous), substrates. The mechanisms of these CVD reactions can be highly complex, often involving interactions with solid surface sites, and the combination of heat and momentum transport coupled with reaction kinetics can lead to complex models (see Figure 3.32). Let us look at some simple, yet important, systems in order to gain a better understanding of the kinetics of CVD processes. Consider the growth of a germanium (Ge) film in the production of a semiconducting thin film [16]. The germanium can be deposited onto a substrate through the reaction of GeCl4 with hydrogen gas: Ge(s) + 2HCl(g) + Cl2 (g)

GeCl4 (g) + H2 (g)

(3.107)

The proposed mechanism for this overall reaction involves a gas-phase dissociation step GeCl2 (g) + Cl2 (g)

GeCl4 (g)

(3.108)

two adsorption processes, where S is a surface site on the solid KA

GeCl2 (g) + S

GeCl2 žS

(3.109)

2HžS

(3.110)

KH

H2 (g) + 2S and a surface reaction ks

GeCl2 žS + 2HžS −−−→ Ge(s) + 2HCl(g) + 2S

Metal Center

(3.111)

Ligand

Metal-containing Molecule + (1) Transport

Gas-phase reaction

(2) Adsorption

(7) Transport

(6) Desorption (4) diffusion

(3) reaction

(5) Nucleation and growth

+ Substrate

Figure 3.32 Processes involved in chemical vapor deposition. From Chemistry of Advanced Materials: An Overview, L. V. Interrante and M. J. Hampden-Smith, eds. Copyright  1998 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

KINETIC PROCESSES IN COMPOSITES

271

The rate-limiting step is believed to be the surface reaction between adsorbed molecular hydrogen and germanium dichloride (3.111): d[Ge] = ks [GeCl2 žS][HžS]2 dt

(3.112)

where the concentration of adsorbed surface species is often represented by their fractional coverage, fi , or the fraction of the surface occupied by species i: d[Ge] = ks fGeCl2 žS fH2 žS dt

(3.113)

The deposition rate, or film growth rate, is usually expressed in nm/s, as is the reaction rate constant, ks . Deposition rates can be expressed in mol/m2 žs by multiplying by the molar density of solid germanium. It can be shown that through appropriate substitution of Eqs. (3.108) through (3.110) and use of the fractional area balance (fv + fGeCl2 + fH = 1, where fv is the fraction of vacant surface sites), the deposition rate can be expressed in terms of the partial pressures (concentrations) of the reactants as k ′ PGeCl2 PH2 d[Ge] = dt (1 + KA PGeCl2 + KH PH2 )3

(3.114)

where k ′ is the product of the two adsorption equilibrium constants, KA and KH , and the reaction rate constant ks . Another example of an industrially relevant CVD process is the production of tungsten (W) films on various substrates for (a) wear and corrosion protection and (b) diffusion barriers in electronic devices. One CVD method for depositing tungsten is through the reduction of WF6 with hydrogen, which follows the overall reaction: WF6 (g) + 3H2 (g)

W(s) + 6HF(g)

(3.115)

This reaction is actually slightly endothermic (H = 88 kJ/mol), but the large net increase in entropy and the nonequilibrium nature of most CVD processes lead to significant tungsten deposition. As with the Ge example, the deposition mechanism involves adsorption steps and surface reactions. At low pressures and under conditions of excess hydrogen gas, the deposition rate follows the general form: d[W(s)] 1/2 0 ∝ PH2 PWF 6 dt

(3.116)

Ideally, the functional dependence of the growth rate on the pressure (concentration) of the gaseous species would be determined by analysis of all the mechanistic steps, as was the case for our Ge example and as could be done for the W example. This is not always possible, however, due to insufficient information on such factors as reaction rate constants, adsorption energies, and preexponential factors. As a result, deposition rate data are often determined experimentally and are used to either support or refute

272

KINETIC PROCESSES IN MATERIALS

900

1100

1300

102

1500

1900 1700

Temperature (°C)

Rate (mils/hr)

101

100

10−1

10−2

4

5

6

7

8

9

1/T (°K −1) × 104

Figure 3.33 Arrhenius plot for the deposition of Si3 N4 . Reprinted, by permission, from F. Galasso, Chemical Vapor Deposited Materials, p. 23. Copyright  by CRC Press.

suspected rate-limiting steps. Arrhenius-type plots are often made by plotting the logarithm of the measured deposition rate versus inverse temperature (see Figure 3.33 for the formation of Si3 N4 from SiF4 and NH3 ), sometimes for various total gas pressures (see Figure 3.34 for the formation of B4 C from BCl3 , CH4 , and H2 ). There are numerous materials, both metallic and ceramic, that are produced via CVD processes, including some exciting new applications such as CVD diamond, but they all involve deposition on some substrate, making them fundamentally composite materials. There are equally numerous modifications to the basic CVD processes, leading to such exotic-sounding processes as vapor-phase epitaxy (VPE), atomiclayer epitaxy (ALE), chemical-beam epitaxy (CBE), plasma-enhanced CVD (PECVD), laser-assisted CVD (LACVD), and metal–organic compound CVD (MOCVD). We will discuss the specifics of CVD processing equipment and more CVD materials in Chapter 7. 3.4.2

Kinetics of Chemical Vapor Infiltration∗

Chemical vapor infiltration (CVI) is similar to CVD in that gaseous reactants are used to form solid products on a substrate, but it is more specialized in that the substrate is generally porous, instead of a more uniform, nominally flat surface, as in CVD. The porous substrate introduces an additional complexity with regard to transport of the reactants to the surface, which can play an important role in the reaction as illustrated earlier with CVD reactions. The reactants can be introduced into the porous substrate by either a diffusive or convective process prior to the deposition step. As infiltration proceeds, the deposit (matrix) becomes thicker, eventually (in the ideal situation) filling the pores and producing a dense composite.

KINETIC PROCESSES IN COMPOSITES

273

Temperature (°C) 1

1200

1100

6.6 kPa Log Deposition Rate (mg cm−2 h−1)

0

1000

900

800

13.2 kPa

2 kPa

−1

−2

−3

−4 0.65

0.75

0.85

0.95

103/T(K−1)

Figure 3.34 Arrhenius plot for the deposition of B4 C at various total pressures. Reprinted, by permission, from R. Naslain, CVI composites, in Ceramic-Matrix Composites, R. Warren, ed., pp. 199–244. Copyright  1992 by Chapman & Hall, London.

There are five general categories of CVI processes, depending on which type of transport process they utilize (diffusion or convection) and whether there is an imposed thermal gradient or not (see Figure 3.35). Type I, known as isothermal CVI (ICVI) has no thermal gradients imposed, and reagents enter the substrate via diffusion. The disadvantage of this process is that the entire substrate is at the reaction temperature, so that pores on the surface can close off before the inner pores are completely filled. This problem is circumvented with a thermal gradient, as in Type II CVI. Here, the reactants contact the substrate in a cold region, and diffuse to the heated reaction zone. Type III and Type IV CVI are the forced convection analogues of Type I and Type II, respectively, where the reactants are now forced into the substrate by convection instead of being allowed to diffuse. Type V CVI is specialized, in that not only are the reactants forced into the substrate, but gaseous products are also forcefully removed in a cyclical manner. The isothermal reaction/diffusion system (Type I) is the easiest to model and provides us with some insight into the competition between kinetics and diffusion. Consider the case of CVI in a filament bundle, as shown schematically in Figure 3.36. The regions between the fibers can be approximated as cylindrical pores with diameter r0 and depth L. Ideally, reactants will diffuse into the pore and create a uniform concentration throughout the pore length before they react to form product, thereby filling the pore uniformly from bottom to top. This is often not the case, and there can be a concentration gradient, as shown in Figure 3.36, leading to nonuniform deposition.

274

KINETIC PROCESSES IN MATERIALS

HISTORICAL HIGHLIGHT

The dream to make synthetic diamond spans over a century. Some of the nineteenth century’s most decorated scientists went to great extremes to coerce carbon-bearing ingredients into the coveted crystalline structure of diamond. In 1880, for example, the Scottish chemist James Ballantine Hannay filled tubes with various carboncontaining ingredients such as burned sugar and bone ash, sealed the tubes, and then heated them until they exploded. He claimed that tiny hard particles retrieved from the debris were diamonds. (He also attributed a nervous disorder he developed to the frequent fear of injury during these explosions.) Later analysis of his experimental protocol virtually precluded the possibility of his actually having produced diamonds. Robert Hazen, a geophysicist at the Carnegie Geophysical Laboratory in Washington, D.C., speculates that Hannay’s

assistants may have spiked his preparations with tiny diamonds with the hope of getting Hannay to give up his mortally dangerous obsession. The first documented successes occurred in 1953 and 1954. The earliest one was quiet and remained unannounced by cautious researchers at Sweden’s main electrical company. The other was announced with fanfare in February of 1955 by the General Electric Company before the Department of Defense silenced the company for several years in the interest of national security. The youngest member of the GE diamondmaking teams, the freshly graduated Robert Wentorf, Jr., even found that he could convert peanut butter, a rather popular carbon-containing material in its own right, into diamond grit. Source: Stuff, I. Amato, p. 139.

Hot

Hot

Cold Type II Thermal gradient: Reagents contact cold surface of preform and enter via diffusion

Hot Type I Isothermal: Reagents surround preform and enter via diffusion (conventional CVI:ICVI) Hot

Hot

Hot

Cold

Type III Isothermal-forced Type IV Thermal gradientflow: reagents flow forced flow: through preform reagents flow through preform from cold to hot surface (forced CVI:FCVI)

Type V Pulsed flow: reagents flow into and out of preform because of cyclical evaluation and back filling

Figure 3.35 Classification of chemical vapor infiltration processes. From Carbide, Nitride, and Boride Materials Synthesis and Processing, A. W. Weimer, ed. p. 563. Copyright  1997 by Chapman & Hall, London, UK, with kind permission of Kluwer Academic Publishers.

KINETIC PROCESSES IN COMPOSITES

275

Filaments r0 Filament Bundle C0

C0

Deposited matrix

r0 Pore

Pore depth L

Reactant

Reactant concentration C

Figure 3.36 Schematic diagram of pore-filling process in the chemical vapor infiltration of a fiber bundle. From Carbide, Nitride, and Boride Materials Synthesis and Processing, A. W. Weimer, ed., p. 566, Copyright  1997 by Chapman & Hall, London, with kind permission of Kluwer Academic Publishers.

Eventually, the top of the pore will close off, and reactants will not be able to find their way to the bottom of the pore to complete the densification process. The competition between kinetics and transport processes can be characterized using one of two dimensionless groups. The first is called the Dahmk¨ohler number, Da, k1 r0 /D, where k1 is the reaction rate constant for the rate-limiting step, r0 is the pore radius, and D is the diffusivity of the gaseous reactants. A related dimensionless group is the Thiele modulus, 2 = 2k1 L2 /r0 D, where the additional parameter, L, is the pore length. Both groups represent a ratio of the kinetic deposition rate to the gas phase mass transport rate, or, similarly, the diffusion time to the reaction time. For very small values of either dimensionless group, gas transport is very rapid and the reaction rate is very slow. This is the ideal situation, with uniform deposition along the pore length the result. In contrast, deposition is rapid for large values of Da and 2 , but diffusion is slow, so there is a decreasing concentration of reactants down the pore length and a nonuniform density results. Mixed control occurs when the reaction and transport rates are comparable (Da ≈ 2 ≈1). The CVI procedure can lead to some very interesting composite materials. Lackey et al. [17] have produced composites of carbon fibers in a laminated SiC/carbon matrix (see Figure 3.37). The alternating SiC/carbon layers were produced by CVI of two different reactions: the decomposition of methyltrichlorosilane (MTS) in the presence of hydrogen; and the decomposition of propylene in hydrogen, respectively. There is also an excellent opportunity to model these processes with the intent at gaining better control of the CVI process [18].

276

KINETIC PROCESSES IN MATERIALS

Cooperative Learning Exercise 3.5

Consider the thermal decomposition of trichloromethylsilane (TMS) in the presence of hydrogen gas to give SiC in a porous substrate: k1

(CH3 )SiCl3 (g) + H2 (g) −−−→ SiC(s) + 3HCl(g) + H2 (g) The following operating conditions are typically used for this CVI process: Pore diameter Pore length Ambient temperature Inlet pressure Molecular diffusivity Preexponential reaction constant Activation energy

1–10 µm 100–1000 µm 800–1200◦ C 5–200 torr 726 cm2 /s@1000◦ C and 5 torr 262 cm/s 120 kJ/mol

Person 1: Estimate the Dahmk¨ohler number, Da. Person 2: Estimate the Thiele modulus, 2 . Compare your results. Is this CVI process reaction-controlled, diffusion-controlled, or a combination of both? Answer : For T = 1273 K, r0 = 0.0005 cm, and L = 0.1 cm, Da is about 10−8 , and 2 is about 10−4 , both of which indicate a reaction-controlled CVI process. Figure 3.37 CVI-laminated SiC (light) and carbon (dark) matrix layers surrounding carbon fibers. Reprinted, by permission, from W. J. Lackey, S. Vaidyaraman, and K. L. More, J. Am. Ceram. Soc., 80, 113. Copyright  1997 by The American Ceramic Society.

KINETIC PROCESSES IN BIOLOGICS

3.5

277

KINETIC PROCESSES IN BIOLOGICS∗

To attempt to describe the kinetics of formation, adaptation, and decomposition of tissue is to try and detail the processes inherent to life itself. This presents a rather daunting challenge, even if we limit ourselves to human tissue. Luckily, other disciplines such as biology have performed this task for us. The factors involved in the reproduction and growth of living tissue, as well as the kinetics of energy production, respiration, and excretion in cells, are not of concern to us here. What we are hoping to understand are the kinetic processes that will eventually allow us to seamlessly integrate biological and synthetic systems. We aim to describe not only the building blocks of biological systems (as we did in Chapter 1), and whether these building blocks are thermodynamically compatible with each other (as we did in Chapter 2), but also how rapidly the incorporation of synthetic materials into living systems can proceed. We will describe the kinetics of one such process here, namely, the kinetics of receptor–ligand binding in cells. 3.5.1

Kinetics of Cell Surface Receptor–Ligand Binding∗

Recall from Sections 1.5 and 2.5 that there are types of specific binding in and between cells called receptor–ligand binding (see Figure 3.38). The receptor–ligand interaction phenomenon is an extremely important one—one that goes far beyond simple cell adhesion. Such other cellular functions as growth, secretion, contraction, and motion are known to be receptor-mediated cell behaviors. Moreover, an understanding of the receptor–ligand interaction affords us the potential to control these cellular functions. For example, cell biologists can chemically alter the structure of the receptor and/or the ligand in order to modify the association and dissociation rate constants. Figure 3.38 suggests that we can write a reversible reaction for the formation of the receptor–ligand complex kf

R+L

C

(3.117)

kr

and that the time rate of change of the complex density is given by d[C] = kf [R][L] − kr [C] dt

(3.118)

Let us continue by considering the simplest case, where the ligand concentration remains essentially constant at its initial value of [L]0 , and the concentration of unbound

+

R

L

C

Figure 3.38 Schematic representation of receptor (R)–ligand (L) binding to form a receptor/ligand complex (C).

278

KINETIC PROCESSES IN MATERIALS

receptors is given by the difference between the total concentration of receptors, [R]T (assumed to be constant), and the concentration of receptors in the complex, [RT − C]: d[C] = kf [RT − C][L]0 − kr [C] dt

(3.119)

The use of dimensionless groups will simplify the solution to Eq. (3.119). Let the concentration of the complex, [C], be scaled to the total receptor concentration, such that u = [C]/[R]T and the dimensionless time, τ , is given by τ = kr /t. The ratio of the reverse to forward rate constants is simply the dissociation equilibrium constant (the inverse of our “normal” equilibrium constant), KD = kr /kf , so that Eq. (3.119) now becomes (1 − u)[L]0 du −u (3.120) = dτ KD which we can solve using the boundary condition that u = u0 at τ = 0 to give  

    L0 L0 KD L0 τ + 1 − exp − 1 + τ u(t) = u0 exp − 1 + L0 KD KD 1+ KD

(3.121)

Though this equation seems a bit unwieldy, it is a rather straightforward description of the variation of the scaled complex concentration as a function of time for various L0 /KD ratios. For example, if u0 = 0, as is the case for the beginning of the binding procedure where there is no complex yet formed, a series of plots results as shown in Figure 3.39 for various L0 /KD ratios. Such a plot is general for different types of receptor–ligand pairs. Some typical values of the rate constants and receptor concentrations for investigated receptor–ligand pairs are shown in Table 3.11.

1 L0/KD=10

0.8

u

0.6 1 0.4 0.2 10-1 0 0

0.2

0.4

0.6

0.8

1

τ

Figure 3.39 The change in dimensionless receptor/ligand binding complex concentration, u, with dimensionless time, τ , for various values of L0 /KD . Reprinted, by permission, from D. A. Lauffenburger and J. J. Linderman, Receptors, p. 21. Copyright  1993 by Oxford University Press, Inc.

279

Transferrin 2.4G2 Fab FNLLP Human interferon-α2 a TNF Hydroxybenzylpindolol Prazosin Insulin EGF Fibronectin IgE IL-2

Ligand HepG2 Mouse macrophage Rabbit neutrophil A549 A549 Turkey erythrocyte BC3H1 Rat fat cells Fetal rat lung Fibroblasts Human basophils T lymphocytes

Cell Type

kf (M min−1 ) 3 × 106 3 × 106 2 × 107 2.2 × 108 9.6 × 108 8 × 108 2.4 × 108 9.6 × 106 1.8 × 108 7 × 105 3.1 × 106 2.3 × 107 8.4 × 108 1.9 × 109

5 × 104 7.1 × 105 5 × 104 900 6.6 × 103 — 1.4 × 104 1 × 105 2.5 × 104 5 × 105 — 2 × 103 1.1 × 104 2 × 103

−1

RT (#/cell)

0.1 0.0023 0.4 0.072 0.14 0.08 0.018 0.2 0.12 0.6 0.0015 0.015 24 0.014

kr (min−1 )

3.3 × 10−8 7.7 × 10−10 2 × 10−8 3.3 × 10−10 1.5 × 10−10 1 × 10−10 7.5 × 10−11 2.1 × 10−8 6.7 × 10−10 8.6 × 10−7 4.8 × 10−10 6.5 × 10−10 2.9 × 10−8 7.4 × 10−12

KD (M)

HepG2 = human hepatoma cell line; 2.4G2 Fab = Fab portion of 2.4G2 antibody against receptor; FNLLP = N -formylnorleucylleucylphenylalanine; A549 = human lung alveolar carcinoma; TNF = tumor necrosis factor; hydroxybenzylpindolol is an antagonist to the receptor; EGF = epidermal growth factor; IgE = immunoglobulinE; IL-2 = interleukin 2; prazosin is an antagonist to the receptor; BC3H1 = smooth muscle-like cell line; RBL = rat basophilic leukemia cell line. Source: D. A. Lauffenburger and J. J. Linderman, Receptors, p. 30. Copyright  1993 by Oxford University Press, Inc.

Transferrin Fcγ Chemotactic peptide Interferon TNF β-adrenergic α1 -adrenergic Insulin EGF Fibronectin Fcg IL-2 (heavy chain) IL-2 (light chain) IL-2 (heterodimer)

Receptor

Table 3.11 Some Receptor–Ligand Binding Parameters

280

KINETIC PROCESSES IN MATERIALS

REFERENCES Cited References 1. Avrami, M., Kinetics of phase change I: General theory, J. Chem. Phys., 7, 1103 (1939). 2. Avrami, M., Kinetics of phase change II: Transformation–time relations for random distribution of nuclei, J. Chem. Phys., 8, 212 (1940). 3. Avrami, M., Kinetics of phase change III: Granulation, phase change, and microstructure, J. Chem. Phys., 9, 177 (1941). 4. Matusita, K., and S. Sakka, Kinetics study of the crystallization of glass by differential scanning calorimetry, Phys. Chem. Glasses, 20, 81 (1979). 5. Kissinger, H. E., Reactive kinetics in differential thermal analysis, Anal. Chem., 29, 1702 (1957). 6. Baz, A., and T. Chen, Torsional stiffness of nitinol-reinforced composite drive shafts, Composites Eng., 3, 1119 (1993). 7. http://www.seas.upenn.edu/∼zshinar/researchb.html 8. http://aml.seas.ucla.edu/ 9. Lewis, T. H., Take a deeper look at cathodic protection, Chem. Eng. Prog., 95(6), 55 (1999). 10. JANAF Thermochemical Tables, D. R. Stull, and H. Prophet, project directors, U.S. National Bureau of Standards, Washington, D.C., 1971. 11. Parks, G. A., Chem. Rev., 65, 177 (1965). 12. Hogenson, D. K., and J. H. Healy, Mathematical treatment of glass corrosion data, J. Am. Cer. Soc., 45(4), 178 (1962). 13. Vanderbilt, D. P., J. P. English, G. L. Fleming, F. W. McNeely, D. R. Cowsar, and R. L. Dunn, Biodegradable polyamides based on 4, 4′ -spirobibutyrolactone, in Progress in Biomedical Polymers, C. G. Gebelein and R. L. Dunn, eds., Plenum Press, New York, 1990, p. 249. 14. Feijen, J., Biodegradable polymers for medical purposes, in Polymeric Biomaterials, E. Piskin and A. S. Hoffman, eds., NATO ASI Series, Martinus Nijhoff Publishers, Dordrecht, 1986, p. 62. 15. Shih. C., Pharmaceut. Res., 12, 2036 (1995). 16. Ishii, H., and Y. Takahashi, J. Electrochem. Soc., 135, 1539 (1988). 17. Lackey, W. J., S. Vaidyaraman, and K. L. More, Laminated C–SiC matrix composites produced by CVI, J. Am. Ceram. Soc., 80, 113 (1997). 18. Gupte, S. M., and J. A. Tsamopoulos, Densification of porous materials by chemical vapor infiltration, J. Electrochem. Soc., 136, 2–555 (1989).

General Gaskell, David R., Introduction to the Thermodynamics of Materials, 3rd ed., Taylor and Francis, Washington, D.C., 1995. Hill, C. G., An Introduction to Chemical Engineering Kinetics and Reactor Design, John Wiley & Sons, New York, 1977. Fogler, H. S., Elements of Chemical Reaction Engineering, 3rd ed., Prentice-Hall, Upper Saddle River, NJ, 1999. Schmidt, L. D., The Engineering of Chemical Reactions, Oxford University Press, New York, 1998.

REFERENCES

281

Metal Formation Engineering Metallurgy, F. T. Sisco, ed., Pitman Publishing, New York, 1957.

Metal Phase Transformations Haasen, P., Physical Metallurgy, Cambridge University Press, London, 1978. Machlin, E. S., Thermodynamics and Kinetics Relevant to Materials Science, Giro Press, New York, 1991.

Metal Degradation Craig, B. D., Fundamental Aspects of Corrosion Films in Corrosion Science, Plenum Press, New York, 1991. Fontana, M. G., Corrosion Engineering, 3rd edition, McGraw-Hill, New York, 1986. Corrosion Mechanisms in Theory and Practice, P. Marcus, edi. Marcel Dekker, New York, 2002.

Ceramic Formation Weimer, A. W., Carbide, Nitride and Boride Materials Synthesis and Processing, Chapman Hall, London, 1997.

Ceramic Phase Transformations Kingery, W. D., H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics, John Wiley & Sons, New York, 1976.

Ceramic Degradation McCauley, R. A., Corrosion of Ceramics, Marcel Dekker, New York, 1995.

Polymer Formation Billmeyer, F. W., Jr., Textbook of Polymer Science, 3rd edition, John Wiley & Sons, New York, 1984. Rodriguez, F., Principles of Polymer Systems, 2nd ed., McGraw-Hill, New York, 1982.

Polymer Degradation Reich, L., and S. S. Stivala, Elements of Polymer Degradation, McGraw-Hill, New York, 1971. Polymers of Biological and Biomedical Significance, S. Shalaby, Y. Ikada, R. Langer, and J. Williams, eds., ACS Symposium Series 540, ACS, Washington, D.C., 1994. Progress in Biomedical Polymers, C. G. Gebelein and R. L. Dunn, eds., Plenum Press, New York, 1988.

CVD/CVI Interrante, L. V., and M. J., Hampden-Smith, Chemistry of Advanced Materials: An Overview, Wiley-CVH, New York, 1998. Galasso, F. S., Chemical Vapor Deposited Materials, CRC Press, Boca Raton, FL, 1991.

282

KINETIC PROCESSES IN MATERIALS

Cell-Surface Receptor–Ligand Binding Lauffenberger, D. A., and J. J. Linderman, Receptors: Models for Binding, Trafficking and Signaling, Oxford University Press, New York, 1993.

PROBLEMS Level I

3.I.1 One-half of an electrochemical cell consists of a pure nickel electrode in a solution of Ni2+ ions; the other is a cadmium electrode immersed in a Cd2+ solution. If the cell is a standard one, write the spontaneous overall reaction and calculate the voltage that is generated. 3.I.2 A piece of corroded steel plate was found in a submerged ocean vessel. It was estimated that the original area of the plate was 64.5 cm2 and that approximately 2.6 kg had corroded away during the submersion. Assuming a corrosion penetration rate of 200 mpy for this alloy in seawater, estimate the time of submersion in years. 3.I.3 For the following pairs of alloys that are coupled in seawater, predict the possibility of corrosion; and if corrosion is possible, note which alloy will corrode: (a) Al/Mg; (b) Zn/low carbon steel; (c) brass/Monel; (d) titanium/304 stainless steel; and (e) cast iron/315 stainless steel. Clearly state any assumptions you make about compositions of alloys. 3.I.4 It is known that the kinetics of recrystallization for some alloy obeys the JMA equation and that the value of n is 2.5. If, at some temperature, the fraction recrystallized is 0.40 after 200 minutes, determine the rate of recrystallization at this temperature. 3.I.5 Consider the condensation polyesterification reaction between ethylene glycol, HO–(CH2 )2 –OH, and terephthalic acid, HOOC–Ph–COOH, each of which has an initial concentration of 1.0 mol/liter. Calculate the number average and weight average degrees of polymerization at 1, 5, and 20 hours. The forward reaction rate constant for the polymerization reaction is 10.0 liter/molžhr, and second-order, catalyzed kinetics can be assumed. Level II

3.II.1 Compute the cell potential at 25◦ C for the electrochemical cell in Problem 3.I.1 if the Cd2+ and Ni2+ concentrations are 0.5 and 10−3 M; respectively. Is the spontaneous reaction direction still the same as for the standard cell? 3.II.2 The corrosion potential of iron immersed in deaerated acidic solution of pH = 3 is −0.720 V as measured at 25◦ C with respect to the normal calomel electrode. Calculate the corrosion rate of iron in millimeters per year, assuming the Tafel slope of the cathodic polarization curve, βc , equals 0.1 V/decade and the hydrogen ion exchange current, i0 , is 0.1 × 10−6 A/cm2 . 3.II.3 In an emulsion polymerization, all the ingredients are charged at time t = 0. The time to convert various amounts of monomer to polymer is indicated in

PROBLEMS

283

the following table. Predict the time for 30% conversion. Polymerization is “normal”—that is, no water solubility or polymer precipitation. Fraction of Original Monomer Charge Converted to Polymer

Time (hours)

0.05 0.12 0.155

2.2 4.0 4.9

3.II.4 Construct a plot analogous to Figure 3.39 for the dissociation of the receptor–ligand complex; that is, for u0 = 1. Make plots for L0 /KD = 10, 1, and 0.1. Level III

3.III.1 As a pure liquid is supercooled below the equilibrium freezing temperature (Tm ), the free energy difference between the solid and liquid becomes increasingly negative. The variation of Gv with temperature is given approximately by Gv = (Hm /Tm )(Tm − T ), where Hm (< 0) is the latent heat of solidification. Using this expression, show that exp(−No G∗ /RT ) is a maximum when T = Tm /3. Why does the maximum nucleation rate occur at a temperature greater than Tm /3? 3.III.2 A review article on the CVD processes used to form SiC and Si3 N4 by one of the pioneers in this area, Erich Fitzer [Fitzer, E., and D. Hegen, Chemical vapor deposition of silicon carbide and silicon nitride—Chemistry’s contribution to modern silicon ceramics, Angew. Chem. Int. Ed. Engl., 18, 295 (1979)], describes the reaction kinetics of the gas-phase formation of these two technical ceramics in various reactor arrangements (hot wall, cold

2

Si3N4 Hot wall (SiHCl3/NH3)

Si3N4 Hot wall (SiCl4/NH3)

log kges [s−1]

1

SiC Cold wall (CH3SiCl3)

0 SiC Fluidized bed (CH3SiCl3)

−1

0.5

0.6

0.7 1/T [K−1]

0.8

0.9

284

KINETIC PROCESSES IN MATERIALS

wall, and fluidized bed) for several precursor combinations, as indicated in the accompanying diagram. (a) Determine the activation energy(ies) for each ceramic/reactor/precursor type. (b) Comment on what causes the break in two of the curves at high temperatures. 3.III.3

Consider three 10-m3 reactors operating at 70◦ C. Each contains 1000 kg of styrene monomer. Additional ingredients initially are as follows: Reactor I

Reactor II

Reactor III

Inert liquid Initiator

8.00 m3 benzene 1.00 kg benzoyl peroxide

8.00 m3 benzene 1.00 kg benzoyl peroxide

Polystyrene

10.0 kg polymer in solution

10.0 kg polymer dissolved in the monomer 1.00 kg poly(vinyl alcohol)

8.00 m3 water 1.00 kg potassium persulfate 10.0 kg polymer in stable latex form —

Additive



In each of the three reactors, does −d[M]/dt (initial value) increase, decrease, or remain essentially the same when each of the following changes is made? (a) Amount of initiator is doubled; (b) amount of monomer is halved; (c) temperature is increased to 90◦ C; (d) amount of initial polymer is doubled. Also, predict the effect of each change on the x n of the polymer being formed initially.

CHAPTER 4

Transport Properties of Materials

4.0

INTRODUCTION AND OBJECTIVES

Perhaps the single largest subject area that differentiates engineering from the sciences is that of transport phenomena. For those of you who have had little or no background in transport processes, you may find some supplemental reading to be of great help in this chapter. There are many excellent textbooks and references on transport phenomena, including the landmark text by Bird, Stewart, and Lightfoot, which first brought the analogies between momentum, heat, and mass transfer to engineering students in a rigorous fashion. We will rely heavily on these analogies in this chapter; and without going into any great depth of coverage on any one transport topic, we will see how transport properties of liquid and solid materials go a long way towards describing many important phenomena in materials selection, design, and processing. The analogy to which we refer is fundamentally quite simple. It states that the rate of flow per unit area, called a flux, of a fundamental quantity such as momentum, heat, or mass is proportional to a corresponding driving force which causes the flow to occur. Flux = “constant” × Driving force (4.1) In this chapter, we will focus mostly on the “constant” in Eq. (4.1), and not the flux or driving force, since these are specific to a process. The proportionality constant is a material property, however, which we can measure, correlate with process parameters, and hopefully predict. The term “constant” is in quotations, because we will see that in some instances, it is really a function of the driving force. It is important to understand (a) the driving forces under which this proportionality constant operates and (b) the effect the constant has on transport phenomena. The transport process about which most of us have an intuitive understanding is heat transfer so we will begin there. In order for heat to flow (from hot to cold), there must be a driving force, namely, a temperature gradient. The heat flow per unit area (Q/A) in one direction, say the y direction, is the heat flux, qy . The temperature difference per unit length for an infinitesimally small unit is the temperature gradient, dT /dy. According to Eq. (4.1), there is then a proportionality constant that relates these two quantities, which we call the thermal conductivity, k. Do not confuse this quantity with An Introduction to Materials Engineering and Science: For Chemical and Materials Engineers, by Brian S. Mitchell ISBN 0-471-43623-2 Copyright  2004 John Wiley & Sons, Inc.

285

286

TRANSPORT PROPERTIES OF MATERIALS

rate constants of the previous chapter, or with the Boltzmann constant, kB . The form of Eq. (4.1) for heat transfer, then, is called Fourier’s Law: qy = −k

dT dy

(4.2)

The minus sign in Eq. (4.2) arises due to the fact that in order for there to be heat flow in the +y direction, the temperature gradient in that direction must be negative—that is, lower temperature in the direction of heat flow. If the temperature gradient is expressed in units of K/m, and the heat flux is in J/m2 · s, then the thermal conductivity has units of J/K·m·s, or W/m·K. A related quantity is the thermal diffusivity, which is often represented by the lowercase Greek letter alpha, α. Thermal diffusivity is defined as k/ρCp , where k is the thermal conductivity, ρ is the density, and Cp is the heat capacity at constant pressure per unit mass. We will see in a moment why the term “diffusivity” is used to describe this parameter. We will generally confine our descriptions in this chapter to thermal conductivity. The equations for one-dimensional momentum and mass flow are directly analogous to Fourier’s Law. A velocity gradient, dvx /dy, is the driving force for the bulk flow of momentum, or momentum flux, which we call the shear stress (shear force per unit area), τyx . This leads to Newton’s Law of Viscosity: τyx = −µ

dvx dy

(4.3)

The proportionality constant, µ, is the viscosity. When the velocity gradient is cm/s·cm, or simply s−1 , we call it the shear rate because of the units of inverse time. Often, the lowercase Greek letter gamma with an overdot, γ˙ , is used to represent the shear rate, dvx /dy. The shear stress of x-directed moment in the y direction, τyx , typically has units of g/cm·s2 . Hence, viscosity has cgs units of g/cm·s, or poise. In the SI system, the unit of Pa·s is often used. Note that 1 Pa·s = 10 poise. In this chapter we will have a lot to say about viscosity and how it varies for different materials in their molten or liquid states. Finally, if we have a concentration gradient of one component (A) in a binary mixture (A + B), dcA /dy, this leads to the bulk motion of molecules, known as the molar ∗ flux of component A in the y direction, JAy , which can be related to the concentration gradient using Fick’s Law: dcA ∗ JAy (4.4) = −DAB dy The proportionality constant for mass transport is called the molecular diffusivity of A in B, and it has cgs units of cm2 /s when concentration is mol/cm3 and distance is in centimeters. As a result, the molar flux has units of mol/cm2 ·s. You should now recognize why the term “thermal diffusivity” is used in certain versions of Eq. (4.2). You should be able to see the similarities in Eqs. (4.2) through (4.4). We will use these relationships a great deal in this chapter, so you should understand them fully and do some background reading if they are unfamiliar to you, but we will again concentrate on the proportionality constants: thermal conductivity, viscosity, and molecular diffusivity.

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

287

By the end of this chapter, you should be able to: ž ž ž ž ž ž ž ž

Locate or calculate thermal conductivities, heat capacities, and specific heats for any material. Locate or calculate viscosities for any liquid or molten material. Locate or calculate the binary-component diffusivity or self-diffusion coefficient for a gas, dilute solute, or melt. Estimate the diffusivity of a gas in a porous solid. Differentiate between Newtonian and non-Newtonian fluids, and know which behavior is appropriate for a certain type of material in the fluid state. Convert between thermal conductivity, thermal conductance, thermal resistivity, and thermal resistance for a material of specific geometry. Calculate permeabilities for substances, given diffusivity and solubility data. Calculate partition coefficients across membranes.

Finally, you will notice that there are a few things that are different in this chapter. It is primarily organized by the type of transport processes, then secondarily organized by material class—unlike the previous chapters, which were primarily organized by material type alone. This is done to add continuity to the transport subtopics while maintaining the analogies between them within this chapter. Also, much of the material here is indicated as “optional”; that is, the information on momentum and mass transport properties are not taught to any great extent in a “traditional” materials science-oriented course. This doesn’t mean they aren’t important, even to an introductory course in materials. In this case, it means that they are perhaps best read in conjunction with the chapter on materials processing (Chapter 7), since the driving forces arise due to the type of processing equipment being used—for example, e.g., casting, extrusion, or CVD processing. In any case, as engineers, we need to keep the similarities between the types of transport processes in mind as we study them, regardless of the relative extent to which each one is covered.

4.1

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

In this section, we will examine the microscopic origins of viscosity for each material type, and we will see how viscosity is much more complex than simply serving as a proportionality constant in Eq. (4.3). Ultimately, we will find that viscosity is not a constant at all, but a complex function of temperature, shear rate, and composition, among other things. 4.1.1 Momentum Transport Properties of Metals and Alloys: Inviscid Systems

Most metallic systems above their melting point have very low viscosities, typically less than about 100 poise. Liquids with viscosities below this level are generally termed inviscid, that is, they are water-like in their fluidity. This is a relative statement, of course, since their viscosities are distinctly nonzero, but for all practical purposes the pressure required to cause these fluids to flow or overcome surface tension forces,

288

TRANSPORT PROPERTIES OF MATERIALS

for example, is vanishingly small. Nevertheless, the relative simplicity of metallic melts allows us to explore, at least briefly, the molecular origins of viscosity. We will then explore the temperature and composition dependences of metal and alloy viscosities. 4.1.1.1 The Molecular Origins of Viscosity. Viscosity, like all of the transport properties in this chapter, has its roots in the kinetic theory of gases. Though we concentrate on condensed phases in this book, it is important to look at the viscosity of gases so that we may progress to the more complex topic of liquid viscosities, and so that we might see the analogies with the molecular origins of thermal conductivity and molecular diffusion in subsequent sections. We already have much of the information we need to derive an expression for the viscosity of a gas. Recall from Chapter 1 that the atoms or molecules of a gas interact with one another. This interaction can be described in terms of repulsive and attractive energies, as given by Eq. (1.11). Recall also that the force, F , can be related to the energy, U , by Eq. (1.13). A potential energy function that has been extensively studied is the Lennard-Jones potential of Eq. (1.12). We can modify the form of Eq. (1.12) slightly to give the following potential energy function

   σ 12  σ 6 − U (r) = 4ε r r

(4.5)

where σ is the collision diameter of the molecule and ε is the maximum energy of attraction between a pair of molecules. These two parameters are known for many substances and can be approximated for others using fluid critical properties, boiling point, or melting point. From this potential energy function, a relationship for viscosity (in poise) can be derived for a gas at low density: −5

µ = 2.6693 × 10



MT σ 2

(4.6)

where T is the temperature of the gas in degrees Kelvin, M is the molecular weight ˚ and  is a dimensionless quantity that correlates with the of the gas, σ is in A, dimensionless temperature, kB T /ε. We will see that the dimensionless temperature plays a role in thermal conductivity and mass diffusivity as well. For rigid spheres (no attraction or repulsion),  = 1, so that it represents a deviation of the real molecules from a rigid-sphere model. Note that the viscosity of gases at low density increases with temperature and that there is no dependence on pressure in this regime. So, we see that, in theory, there is indeed a link between the potential energy function of a gas and its viscosity. For liquids, even simple monatomic liquids such as molten metals, this link between molecular interactions and a bulk property such as viscosity is still there, in principle, but it is difficult to derive a relationship from fundamental interaction energies that is useful for a wide variety of systems. Nonetheless, theoretical expressions for liquid viscosities do exist. Some are based on statistical mechanical arguments, but the model that is most consistent with the discussion so far is that of nonattracting hard spheres

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

289

in a dense fluid. The viscosity (in poise) of such a liquid is

µ = 3.8 × 10−8



PF PF 1− MT  2  V 2/3  (1 − PF )3





4/3

   

(4.7)

where M, T are the molecular weight and temperature as before, V is the molar volume of the liquid, and PF is the packing fraction or packing density of spheres in the liquid, which can be related to the collision diameter, σ , as PF =

πn0 σ 3 6

(4.8)

where n0 is the average number density (N/V ). Some viscosities for elemental liquid metals as calculated from Eq. (4.7) are compared with experimental values in Table 4.1. Notice the similarity between the relationship for liquid viscosity [Eq. (4.7)] and that for gaseous viscosity [Eq. (4.6)]. They both have a square root dependence on temperature and molecular weight and depend on the inverse square of the collision diameter [can you prove this for Eq. 4.7)?]. So, at least in principle, there is a fundamental relationship between the structure of a liquid and its viscosity. Table 4.1 Comparison of Calculated [using Eq. (4.7) with PF = 0.45] and Experimental Values of Liquid Metal Viscosities Near Their Melting Points Metal Na Mg Al K Fe Co Ni Cu Zn Ga Ag In Sn Sb Au Hg Tl Pb Bi

µcal (mPa·s)

µexp (mPa·s)

0.62 1.69 1.79 0.50 4.55 4.76 4.76 4.20 2.63 1.63 4.07 1.97 2.11 2.68 5.80 2.06 2.85 2.78 2.54

0.70 1.25 1.2–4.2 0.54 6.92 4.1–5.3 4.5–6.4 4.34 3.50 1.94 4.28 1.80 1.81 1.43 5.38 2.04 2.64 2.61 1.63

Source: T. Lida and R. I. L. Guthrie, The Physical Properties of Liquid Metals. Copyright  1988 Oxford University Press.

290

TRANSPORT PROPERTIES OF MATERIALS

4.1.1.2 Temperature Dependence of Pure Metal Viscosity. Practically speaking, empirical and semiempirical relationships do a much better job of correlating viscosity with useful parameters such as temperature than do equations like (4.7). There are numerous models and their resulting equations that can be used for this purpose, and the interested student is referred to the many excellent references listed at the end of this chapter. A useful empirical relationship that we have already studied, and that is applicable to viscosity, is an Arrhenius-type relationship. For viscosity, this is  Hµ µ = A exp (4.9) RT

where A and Hµ are constants. There are several methods for evaluating these two constants, but the simplest expressions relate them to the melting point of the metal, Tm . (We know that this makes some fundamental sense, because the melting point is related to the intermolecular attraction between atoms.) The preexponential factor is given by √ 5.7 × 10−2 MTm A(cpoise) = 2/3 (4.10) Vm exp(Hµ /RTm ) The activation energy for viscous flow, Hµ , for “normal” metals is Hµ = 1.21Tm1.2

(4.11)

and the activation energy for semi-metals is Hµ = 0.75Tm1.2

(4.12)

The two relationships for activation energy as a function of melting point are shown graphically in Figure 4.1. 4.1.1.3 Composition Dependence of Alloy Viscosity. Attempts have been made to calculate the viscosity of a dilute liquid alloy from a theoretical standpoint, but with little success. This is primarily due to the fact that little is known about the interaction of dissimilar atoms in the liquid state. Empirical relationships for the viscosity of dilute liquid alloys have been developed, but these are generally limited to specific alloy systems—for example, mercury alloys with less than 1% impurities. The viscosities of binary liquid alloys have been empirically described using a quantity called the excess viscosity, µE (not to be confused with the excess chemical potential), which is defined as the difference between the viscosity of the binary mixture (alloy), µA , and the weighted contributions of each component, x1 µ1 and x2 µ2 :

µE = µA − (x1 µ1 + x2 µ2 ) = −2(x1 µ1 + x2 µ2 )

HE RT

(4.13)

where x1 and x2 are the mole (atomic) fractions of each component, and H E is the enthalpy of mixing. This expression indicates that the sign of the excess viscosity (either − or +) depends only on the sign of the enthalpy of mixing, which can also

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

291

15.0 Fe Cu Co Ca Ni Ag Au

Normal metals

10.0

Semi-metals

Hm/kcal mol−1

5.0

Mg Cd Tl Pb

2.0

Zn Sb

Rb Na Li Cs K In SnBi

1.0

Ga Hg

0.5 0.3

Al

1

2

5

10

20

40

Tm / 102K

Figure 4.1 Dependence of activation energy for viscous flow on melting point for “normal” metals and semi-metals. Reprinted, by permission, from T. Iida and R. I. L. Guthrie, The Physical Properties of Liquid Metals, p. 187. Copyright  1988 by Oxford University Press.

be either negative or positive. This relationship is only true some of the time, as borne out by experimental data (see Table 4.2). 4.1.2

Momentum Transport Properties of Ceramics and Glasses

4.1.2.1 Newtonian Viscosity in Glasses. As we saw in Chapter 1, the structure of glasses is fundamentally different from metals. Unlike metals and alloys, which can be modeled as hard spheres, the structural unit in most oxide glasses is a polyhedron, often a tetrahedron. As a result, the response of a structural unit to a shear force is necessarily different in molten glasses than in molten metals. The response is also generally more complicated, such that theoretical descriptions of viscosity must give way completely to empirical expressions. Let us briefly explore how this is so. Consider the melting of a common glass such as vitreous silica, SiO2 . Recall that the structural unit in the silicates is the (SiO4 )4− tetrahedron (see Figure 1.43). As heat is provided to the glass, Si–O bonds begin to rupture, and some structural units become mobile. Upon reaching the melting point, a certain portion of broken bonds is sufficient to cause flow in response to shear, but not all the bonds have been broken. For this reason, the SiO2 melt has a high viscosity. As temperature is increased further above the melting point, more bonds are broken and the viscosity decreases. For silicate glasses, this change in viscosity can cover many order of magnitudes. Thus, viscosity–temperature plots for glasses are often given as semilogarithmic plots, as in Figure 4.2. The large range of viscosities for glasses is related to its structure. For example, compared to a highly mobile liquid like water, the relaxation time is much longer in glasses. Relaxation time is the time needed for the structural components to adjust to

292

TRANSPORT PROPERTIES OF MATERIALS

Table 4.2 Relationship Between Excess Viscosity and Difference in Ionic Radii |d1 − d2 | and Enthalpy of Mixing for Various Binary Liquid Alloys

Alloy Au–Sn Ag–Sb Ag–Sn Al–Cu Au–Cu K–Na Cd–Sb Cu–Sb Ag–Cu Hg–In Cu–Sn Al–Zn Cd–Bi K–Hg Pb–Sb Mg–Pb Na–Hg Al–Mg Cd–Pb Pb–Sn Sb–Bi Ag–Au Fe–Ni In–Bi Mg–Sn In–Pb Sn–Zn Sn–Bi

|d1 − d2 | ˚ (A) 0.66 0.64 0.55 0.46 0.46 0.38 0.35 0.34 0.30 0.29 0.25 0.24 0.23 0.23 0.22 0.19 0.15 0.15 0.13 0.13 0.12 0.11 0.10 0.07 0.06 0.03 0.03 0.03

µE

HE

− − − − 0, − − − − − −, +a − − − + − +, −a + + − 0 − + −, +a + + − − +

− − −, +a − − + +, −a − + − − + + − +, −a − − − + + + − − − − + + +

a

In some instances the signs are composition-dependent. Source: T. Iida and R. I. L. Guthrie, The Physical Properties of Liquid Metals. Copyright  1988 by Oxford University Press.

the deformation forces resulting from viscous flow. In general, this is the same order of magnitude as the time required for a structural unit to move approximately one of its own diameters. The smaller the structural units and the less strongly they interact with each other, the shorter the relaxation time. For water, the relaxation time is of the order 10−12 s. For glasses near their softening points, this relaxation time could be around 102 s. It is interesting to note that almost all glasses have a inflection point in the log µ versus temperature curve at about 1015 Pa·s (1016 poise, just beyond where the data end in Figure 4.2), above which the viscosity begins to level off. This is the point at which brittleness begins to set in. There are a number of named temperature regimes that have been assigned to glasses, within which certain characteristic melt properties, primarily related to

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

293

Temperature (°F) 400 1016 1014

800

1200

1600

Borosilicate 96% silica glass glass

2000

2400

Fused silica

2800

3200 18 10 1016

Strain point Annealing point

1012

1014

1010 108 Softening point

108

106 106

Working range 104 102

Working point Melting point Soda-lime glass

1 200

400

600

Viscosity (P)

Viscosity (Pa-s)

1012 1010

104 102

1 800 1000 1200 1400 1600 1800 Temperature (°C)

Figure 4.2 Dependence of viscosity on temperature for several silica-based glasses. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 425. Copyright  2000 by John Wiley & Sons, Inc.

formability, have been identified. From highest to lowest temperature (as is found in a cooling operation), these temperatures include the melting point (∼10 Pa·s), where the glass is fluid enough to be considered a liquid; the working point (∼103 Pa·s), where the glass is easily deformed; the softening point (∼106 Pa·s), where a glass piece can be handled without significant dimensional change; the annealing point (∼1012 Pa·s), where the relaxation time is on the order 103 s; and the strain point (∼1013 Pa·s), where fracture occurs before plastic deformation in response to an applied force (see Chapter 5 for a comparison of fracture and plastic deformation). Other terminology used in the glass industry includes the fictive temperature, where the properties of the glass are the same as those of the melt in metastable equilibrium; the devitrification temperature, where the glass transforms from a noncrystalline state to one that is crystalline; and the Littleton temperature, where a glass rod heated at 5–10 K/m will deform at a rate of 1 mm/min under its own weight. Most viscosity–temperature relationships for glasses take the form of an Arrhenius expression, as was the case for binary metal alloys. The Vogel–Fulcher–Tammann (VFT) equation is one such relationship.   Eµ µ(T ) = A exp (4.14) R(T − T0 ) where Eµ is the activation energy for viscosity, A is the preexponential factor, R is the gas constant, and T0 is an additional adjustable parameter specific to a glass.

294

TRANSPORT PROPERTIES OF MATERIALS

Cooperative Learning Exercise 4.1

Three-parameter models such as the VFT equation (4.14) can be fit to data by first placing the equation in linear form: ln µ = A +

Eµ R(T − T0 )

then using three pairs of dependent/independent variable values: (ln µ1 , T1 ); (ln µ2 , T2 ), and (ln µ3 , T3 ). With three such pairs of data, you can create three independent equations and solve them simultaneously using simple algebraic substitution or using matrices (see Problem 4.III.2 at the end of the chapter). Alternatively, the parameters can be determined sequentially using the following relations, which are represented for the VFT equation, where Li is shorthand notation for ln µi : A= T0 =

(L1 − L2 )(L3 T3 − L2 T2 ) − (L2 − L3 )(L2 T2 − L1 T1 ) (L1 − L2 )(T3 − T2 ) − (L2 − L3 )(T2 − T1 ) L1 T1 − L2 T2 − A(T1 − T2 ) L1 − L2

Eµ /R = (Ti − T0 )(Li − A)

Use the following viscosity–temperature data to complete the following exercise: ln µ1 = 13.3, Tg = 550◦ C; ln µ2 = 7.6, Littleton temperature = 715◦ C; ln µ3 = 4.0, working point = 1000◦ C. Person 1 : Determine the VFT parameters A, Eµ , and T0 from the data provided by solving the three independent equations created when substituting each data pair into the linearized form of the VFT equation. Person 2 : Determine the VFT parameters A, Eµ , and T0 from the data provided by using the three relationships provided above. Compare your answers! Answer : ln µ = −1.386 + 38030.3/(T − 288.8)

Viscosity–temperature plots can be generated for glasses with three adjustable parameters using the VFT equation. 4.1.2.2 Non-Newtonian Viscosities in Ceramic Slurries and Suspensions. As the melting point of inorganic materials such as oxides, borides, and carbides increases, melt processing becomes less practical due, for example, to increased energy requirements, difficulty in containment, or reduced formability. As a result, high-melting-point ceramics are often processed in the form of slurries, suspensions, or pastes. The distinction between a paste, slurry, or suspension depends upon a number of factors, such as (a) whether the ceramic particles, liquid phase, and trapped gases are continuous or discontinuous and (b) the value of particle parameters such as packing fraction (PF), degree of pore saturation (DPS), and mechanical compressibility (X), as summarized in Table 4.3. In a powder composed of spherical particles, the packing fraction (PF) is the fraction of the bulk volume occupied by spheres. As described in Section 1.1.1, the packing fraction for simple cubic is 0.52, and 0.74 for close-packed structures.

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

295

Table 4.3 Characteristics of Ceramic/Liquid Mixture Consistency Statesa

Consistency

Particle Structure

Gas

Liquid

PF

DPS

X

Powder Granules Plastic Paste Slurry Suspension

C C C C D D

C C D D D D

D D C C C C

L L-M M-H M-L M-L L

1

Figure 4.3 Liquid distribution in a powder leading to varying degree of pore saturation (DPS).

The DPS is a measure of the pore saturation and is defined as the ratio of volume of the suspending liquid to pore volume (see Figure 4.3). The mechanical compressibility of a collection of particulates, X, is defined as ( V /V0 )/ P , where V is the volume change resulting from an incremental pressure change P , and V0 is the initial volume. As more liquid is added to the particles, the mixture becomes essentially incompressible, and X approaches zero. These distinctions are important from a processing standpoint, since the transfer of shear forces within the mixture changes significantly. These differences are shown schematically in Figure 4.4. A dry, bulk powder consists of discrete particles and random agglomerates produced by attractive van der Waals, electrostatic, or magnetic interparticle forces. Here, the flow resistance depends upon random particle attraction and agglomeration, and it tends to decrease with decreasing particle size. This occurs because the surface force relative to the gravitational force increases with decreasing particle size. When a wetting liquid of low viscosity is added to a dry powder, agglomerates form. Agglomerates that are purposefully made in this way are called granules, and the free-flowing material is referred to as being granular. The resistance to shearing flow is still low in these mixtures. On further increase of the liquid content,

Apparent Shear Resistance

TRANSPORT PROPERTIES OF MATERIALS

plastic body

paste powder

296

granules

slurry

Concentration of Liquid in Body

Figure 4.4 Processing consistency produced on mixing a powder with a wetting liquid. From J. S. Reed, Principles of Ceramics Processing, 2nd ed. Copyright  1995 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

nearly all of the powder will be transformed into granules with dry surfaces. At this point, only a small amount of liquid is required to cause nearly total agglomeration into a cohesive mass. The cohesive mass is very resistant to shearing, as indicated by the peak in Figure 4.4. The response of this cohesive mass to shear is plastic-like, so it is often called a plastic body. In the agglomerate, plastic, and paste states, shear stress may be transmitted by particles in contact with one another. From this point on, resistance to shear decreases as more liquid is added, and the liquid phase becomes continuous. The flow behavior of a paste depends on the content of colloidal particles and coagulation forces. A paste becomes a slurry upon further addition of liquid. Now the material may flow under its own weight. A dilute slurry is termed a suspension. In addition to temperature, the viscosity of these mixtures can change dramatically over time, or even with applied shear. Liquids or solutions whose viscosity changes with time or shear rate are said to be non-Newtonian; that is, viscosity can no longer be considered a proportionality constant between the shear stress and the shear rate. In solutions containing large molecules and suspensions contain nonattracting anisometric particles, flow can orient the molecules or particles. This orientation reduces the resistance to shear, and the stress required to increase the shear rate diminishes with increasing shear rate. This behavior is often described by an empirical power law equation that is simply a variation of Eq. (4.3), and the fluid is said to be a power law fluid: τ = K γ˙ n (4.15) Here, K is sometimes referred to as the consistency index and has units that depend on the value of the power law index, n—for example, N·sn /m2 . The power law index is itself dimensionless. Typical values of K and n are listed in Table 4.4. In general, the power law index is independent of both temperature and concentration, although fluids tend to become more Newtonian (n approaches 1.0) as temperature increases and concentration decreases. The consistency factor, however, is more sensitive to temperature and concentration. To correct for temperature, the following relationship is often used: Kθ = K0 mθ /m0 (4.16)

297

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

Table 4.4 Power-Law Parameters for Selected Fluids K (N·sn /m2 )

Fluid 23.3% Illinois yellow clay in water1 0.67% carboxymethylcellulose (CMC) in water1 1.5% CMC in water1 3.0% CMC in water1 33% lime in water1 10% napalm in kerosene1 4% paper pulp in water1 54.3% cement rock in water1 0.5% hydroxymethylcellulose (HMC) in water (293 K)2 2.0% HMC in water (293 K)2 1.0% polyethylene oxide in water (293 K)2 Human blood (293 K)3 Ketchup3 Soybean oil3 Sewage sludge3 (1.7–6.2 wt% sol.) w = wt% solids Orange juice3 Corn syrup (48.4 total solids)3

5.55 0.304 3.13 9.29 7.18 4.28 20.0 2.51 0.84 93.5 0.994 0.00384 33.2 0.04 0.00113w 3.4 1.89 0.0053

n 0.229 0.716 0.554 0.566 0.171 0.520 0.575 0.153 0.509 0.189 0.532 0.890 0.242 1.0 0.817–0.843 0.680 1.0

Source: (1) Bird, R. B., W. E. Stewart, and E. N. Lightfoot, Transport Phenomena, John Wiley & Sons, New York, 1960., (2) Dynamics of Polymeric Liquids, Bird, R. B., R. C. Armstrong, and O. Hassager, Vol. 1, John Wiley & Sons, New York, 1977., (3) Johnson, A. T., Biological Process Engineering, WileyInterscience, New York, 1999.

where Kθ is the consistency index at temperature θ , K0 is the consistency factor at the reference temperature, and µθ and µ0 are the corresponding viscosities of the suspending fluid, which is usually, but not always, water. For a power-law fluid, the viscosity is now a function of shear rate. Often times, a non- Newtonian viscosity is designated with the lowercase Greek letter eta, η, to differentiate it from a Newtonian viscosity, µ: η = K γ˙ n−1

(4.17)

In general, we will use µ to represent Newtonian viscosity, and η to represent nonNewtonian viscosity. Recognize, of course, that in certain limits and under certain conditions, non-Newtonian fluids can behave in a Newtonian manner. We will not attempt to switch between the two representations for viscosity under these circumstances. The value of n in Eqs. (4.15) and (4.17) can be greater than or less than 1.0. For n > 1.0, the viscosity increases with shear rate, in what is a called a shear-thickening fluid. Moderately concentrated suspensions containing large agglomerates and concentrated, deflocculated slurries of ceramic powders behave in this manner. When n < 1.0, a shear-thinning fluid results, and less stress is required to incrementally increase the shear rate. Suspensions containing nonattracting, anisometric particles, along with liquids or solutions containing large molecules, are typically shear thinning, because laminar flow tends to orient the particles and reduces the resistance to shear. We will see in the next section that polymeric materials can be both shearthinning and shear-thickening. These two types of power-law behavior are summarized graphically in the lower half of Figure 4.5. Notice that at low shear rates and over

298

TRANSPORT PROPERTIES OF MATERIALS

Shear Thickening with Yield Stress Bingham Shear Stress

Shear Thinning with Yield Stress Shear Thickening Yield Stress Shear Thinning Newtonian Shear Rate

Figure 4.5 Variation of shear stress with shear rate for different models of flow behavior. From J. S. Reed, Principles of Ceramics Processing, 2nd ed. Copyright  1995 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

limited shear rate regimes, shear-thinning and shear-thickening fluids can behave in a psuedo-Newtonian manner. The top three curves in Figure 4.5 represent the behavior of yet another class of nonNewtonian fluids that exhibit a yield stress (note how their curves begin further up on the shear stress axis than the previous group of curves). In ceramic systems, slurries that contain bonded molecules or strongly attracting particles require a finite stress called the yield stress, τy , to initiate flow. Once the yield stress has been overcome, the fluid can behave in either a Newtonian or non-Newtonian manner. Fluids that possess a yield stress, yet have a constant proportionality between stress and strain (shear rate) above the yield stress, are termed Bingham fluids (or Bingham plastics for organic macromolecules) and obey the general relationship: τ = τy + ηp γ˙

(4.18)

Bingham fluids that are either shear-thinning or shear-thickening above their yield stresses have corresponding power-law expressions incorporated into their viscosity models. Further non-Newtonian types of behavior appear when time is introduced as a variable. Fluids whose viscosity decreases with time when sheared at a constant rate are called thixotropic, whereas fluids whose viscosity increases with time when sheared at a constant rate are termed rheopectic. Thixotropic behavior is represented schematically in Figure 4.6. Note that the two viscosities (slope of shear stress vs. shear rate) ηP 1 and ηP 2 are different, depending upon the time they were sheared at a given rate. With this background of non-Newtonian behavior in hand, let us examine the viscous behavior of suspensions and slurries in ceramic systems. For dilute suspensions on noninteracting spheres in a Newtonian liquid, the viscosity of the suspension, ηS , is greater than the viscosity of the pure liquid medium, ηL . In such cases, a relative viscosity, ηr , is utilized, which is defined as ηS /ηL . For laminar flow, ηr is given by the Einstein equation ηr = 1 + 2.5fpν (4.19)

Shear Stress

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

299

Time1

hp

1

Time2 1 1

hp

2

Shear Rate

Figure 4.6 Thixotropic behavior, in which time at constant shear rate affects viscosity. From J. S. Reed, Principles of Ceramics Processing, 2nd ed. Copyright  1995 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Cooperative Learning Exercise 4.2

Agglomeration in a slurry causes a change in the packing factor at which flow is blocked, fcrv , from 0.6 to 0.45 and causes a change in the hydrodynamic shape factor from 3.5 to 2.5. In both cases, the volume fraction of dispersed particles is 0.40. Person 1 : Calculate the relative viscosity for fcrv , = 0.6 and KH = 3.5. Person 2 : Calculate the relative viscosity for fcrv , = 0.45 and KH = 2.5. Compare your answers. How would you expect these values to change if the shear rate were increased? Answers: ηr (0.6, 3.5) = 4.7; ηr (0.45, 2.5) = 9.0 Increased shear rate would help prevent particle agglomeration, effectively increasing fcrv , and reducing the relative viscosity.

where fpv is the volume fraction of dispersed spheres. A more general relationship for the reduced viscosity allows for nonspherical particles, through the use of an apparent hydrodynamic shape factor, KH : ηr = 1 + KH fpv

(4.20)

In suspensions of particles with an aspect ratio (length to diameter) greater than 1, particle rotation during flow results in a large effective hydrodynamic volume, and KH > 2.5 (see Figure 4.7). At particle volume fractions above about 5–10%, interaction between particles during flow causes the viscosity relationship to deviate from the Einstein equation. In such instances, the reduced viscosity is better described by the following relationship: v ηr = [1 − fpv /fcrv ]−KH fp (4.21) Here, fpv is once again the volume fraction of dispersed spheres, and KH is the apparent hydrodynamic shape factor, but a new parameter, fcrv , is introduced which accounts

300

TRANSPORT PROPERTIES OF MATERIALS

v

r

w

w

Relative Viscosity

Figure 4.7 Effect of hydrodynamic shape factor on effective particle size in a flow field. From J. S. Reed, Principles of Ceramics Processing, 2nd ed. Copyright  1995 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

in

Einste

1 0

Volume Fraction Solids

f vcr

Figure 4.8 Effect of increasing the powder concentration in a slurry on relative viscosity. From J. S. Reed, Principles of Ceramics Processing, 2nd ed. Copyright  1995 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

for the packing factor at which flow is blocked. This parameter can be determined graphically, as in Figure 4.8. The effect of volume fraction of dispersed particles on the relative viscosity for several different apparent hydrodynamic shape factors, as determined by Eq. (4.21), is shown in Figure 4.9. There are numerous other equations and correlations that attempt to predict the viscosity of suspensions and slurries. Some take into account adsorbed binder on the particles, and some attempt to predict such things as yield stress and the effects of time on rheological behavior. In most cases, these fluids behave in a highly non-Newtonian manner, and rheological properties can be predicted only for very narrow temperature and concentration ranges. This description of non-Newtonian behavior in ceramic slurries and suspensions leads us directly into a description of viscosity in polymeric materials, which are oftentimes highly non-Newtonian in their behavior.

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

301

30

KH

20

KH = 2.5 V f CR = 0.6 5 fV CR = 0.7

40

= 2.

Relative Viscosity

50

KH = 2.5 V f CR = 0.5

KH = 5.0 f V CR = 0.5

60

KH = 5.0 V f CR = 0.6

70

10

0 0.0

0.1

0.2 0.3 0.4 Volume Fraction of Particles

0.5

0.6

Figure 4.9 Dependence of relative viscosity on solids concentrations for several values of apparent hydrodynamic shape factors. From J. S. Reed, Principles of Ceramics Processing, 2nd ed. Copyright  1995 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

4.1.3

Momentum Transport Properties of Polymers

Unlike metals, in which momentum transport properties are generally limited to the molten state, and ceramics, in which momentum transport properties are primarily (but not always) described by the solution state, polymers can be found in either the solution or molten state. As a result, many of the principles that have been previously discussed apply to polymers, especially with regard to non-Newtonian behavior. There are, however, a few viscosity-related concepts, exclusive to polymers, that we will describe here. 4.1.3.1 Viscosity of Polymer Solutions. Polymer solutions are very common. Glues, pastes, and paints are just a few examples of commercially available aqueous suspensions of organic macromolecules. Also, recall from Chapter 3 that certain types of polymerization reactions are carried out in solution to assist in heat removal. The resulting polymers are also in solution, and their behavior must be fully understand in order to properly transport them and effect solvent removal, if necessary. Unless they are formed directly in solution, polymers dissolve in two general steps. First, the polymer absorbs solvent to form a swollen gel. The second step involves the dissolution of the gel into the solution, sometimes called a sol. Networked or thermoset

302

TRANSPORT PROPERTIES OF MATERIALS

polymers may only undergo gel formation, if they dissolve at all. Once the gel is in solution, the solution viscosity is generally described as it is for ceramic suspensions and slurries, with the relative viscosity ηr =

η ηs

(4.22)

where η is the true solution viscosity, and ηs is the pure solvent viscosity. Keep in mind that not all polymer solutions are aqueous; that is, many polymers dissolve in nonpolar and/or organic solvents, depending on their chemical structure. Equation 4.22 shows us that the relative viscosity for pure solvents is exactly 1.0. Oftentimes, a Taylor series expansion in polymer concentration, c, is used for the relative viscosity (4.23) ηr = 1 + [η]c + k ′ [η]2 c2 + · · · where [η] is called the intrinsic viscosity of the solution, and k ′ is called the Huggins coefficient. The expansion is often truncated at the second-order term. The intrinsic viscosity has units of reciprocal concentration. Both intrinsic viscosity and the Huggins coefficient are independent of concentration., although they can be functions of shear rate. You should be able to see from Eqs. (4.22) and (4.23) that the following relationship holds for the intrinsic viscosity:  η − ηs (4.24) [η] = lim c→0 cηs

Cooperative Learning Exercise 4.3

We are given the following measurements of relative viscosities of solutions of polyisobutylene (M w = 2.5 × 106 ) in decalin and in a 70% decalin and 30% cyclohexanol mixture at 298 K. Person 1 : Calculate the intrinsic viscosity, [η], for the polyisobutylene sample in decalin. Person 2 : Calculate the intrinsic viscosity, [η], for the polyisobutylene sample in decalincyclohexanol. Compare your answers. Which is the “better” solvent?

Decalin 3

Decalin-Cyclohexanol

c(g/cm )

ηr

c(g/cm3 )

ηr

0.000408 0.000543 0.000815 0.001304 0.001630

1.227 1.311 1.485 1.857 2.125

0.000409 0.000546 0.000818 0.001169 0.001637

1.155 1.210 1.326 1.488 1.740

Source: Adapted from Dynamics of Polymeric Liquids

Answers: Decalin 5.1 × 102 cm3 /g; decalin-cyclohexanol 3.1 × 102 cm3 /g.

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

303

It may seem confusing to introduce yet another type of viscosity. In fact, there are many other types of viscosity (see Table 4.5), but we will concentrate on the intrinsic viscosity for polymer solutions. The intrinsic viscosity is important because it has been found to correlate well with a number of important parameters, including polymer molecular weight and type of solvent. Note that if Eq. (4.23) is truncated to the first two terms, the intrinsic viscosity is the same as the reduced specific viscosity, ηsp , of Table 4.5. Polymer solvents are generally characterized as being either a good solvent or a poor solvent. A good solvent is defined as a solvent in which polymer–solvent interactions are more favorable than polymer–polymer interactions. Conversely, in a poor solvent, polymer–polymer interactions are more favorable than polymer–solvent interactions. So, in a good solvent, the polymer chains will tend to uncoil in order to interact more with the solvent, thereby dissolving more effectively. The Huggins coefficient has values in the range 0.3–0.4 for good solvents, with higher values for poor solvents. Intrinsic viscosity has also been found to correlate with the molecular weight of the polymer (see Figure 4.10) according to the Mark–Houwink relationship: a

[η] = K ′ M v Table 4.5

Summary of Viscometric Terms

Symbol η ηs ηr = η/ηs ηsp = ηr − 1 (ln ηr )/c ηsp /c = ηsr

[h] [cm3·g−1]

(4.25)

Name

Common Units

Solution viscosity Solvent viscosity Relative viscosity Specific viscosity Inherent viscosity Reduced viscosity

Poise or Pa·s Poise or Pa·s Dimensionless Dimensionless dl/g dl/g

C6H12 30°

102

DIB 20°

101 104

105 M

106

Figure 4.10 Intrinsic viscosity-molecular weight relationship for polyisobutylene in diisobutylene (DIB) and cyclohexane. Reprinted, by permission, from G. Strobl, The Physics of Polymers, 2nd ed., p. 295. Copyright  1997 by Springer-Verlag.

304

TRANSPORT PROPERTIES OF MATERIALS

Table 4.6

Selected Values of Parameters for Mark–Houwink Equation

Polymer Cellulose triacetate SBR rubber Natural rubber

Solvent

Temperature (◦ C)

K ′ × 105

a

25 25 30 14.5 30

8.97 54 18.5 119 68

0.90 0.66 0.74 0.50 0.66

25 20 135

23.3 20.0 62

0.75 0.66 0.70

Polyacrylamide

Acetone Benzene Benzene n-Propyl ketone Water

Polyacrylonitrile Poly(dimethylsiloxane) Polyethylene

Dimethyl formamide Toluene Decalin

Polyisobutylene

Benzene Benzene Cyclohexane

24 40 30

107 43 27.6

0.50 0.60 0.69

Poly(methyl methacrylate)

Toluene

25

7.1

0.73

Toluene Toluene Benzene Ethyl n-butyl ketone Tetrahydrofuran

30 30 30 29 20

11.0 10.6 22 92.9 3.63

0.725 0.725 0.65 0.50 0.92

Polystyrene Atactic Isotactic Poly(vinyl acetate) Poly(vinyl chloride)

Source: F. Rodriguez, Principles of Polymer Systems, 2nd ed. Copyright  1982 by McGraw-Hill Book Company.

where K ′ and a are constants that depend on the particular polymer- solvent system, and M v is the viscosity-average molecular weight (cf. Table 1.23). The constant a is often called the Mark–Houwink exponent, and lies in the range of 0.5–0.8 (see Table 4.6), which leads to the result that the viscosity-average molecular weight lies between the number-average and weight-average molecular weights. When a = 1.0, M v = M w . Many other important characteristics of the polymer can be predicted using intrinsic viscosity, including chain stiffness and hydrodynamic coil radius, further emphasizing the importance of intrinsic viscosity. 4.1.3.2 Viscosity of Polymer Melts and Blends. Much regarding the sheardependent viscosity of polymer melts has already been described in the context of cermamic slurries (cf. Section 4.1.2.2). There is an immeasurable amount of additional information on this topic, and much research has been conducted on predicting the rheological properties of polymer melts from their structure using what are termed constitutive equations. A review of these equations, even a perfunctory one, is beyond the scope of this text. However, we know enough about polymeric structure and fluid flow to describe two topics on polymer melts that have direct utility with regard to the processing of polymeric materials. The first deals with the effect of chain entanglements on polymer melt viscosity, and the second has to do with the viscosity of polymer melt blends. A third relevant topic, viscoelasticity, will be reserved until Chapter 5 when mechanical properties and elastic response are described.

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

305

The high viscosities of polymeric melts can be attributed to the high level of intermolecular bonding between polymer chains. In fact, the interchain forces can be stronger than primary bonds in the chain—for example, C–C backbone bonds, to the point that thermal degradation often occurs before vaporization in polymers. This phenomenon is chain-length-dependent; that is, the longer the chain, the more interatomic bonding that can occur. Polymer molecules with strong intermolecular forces such as hydrogen bonding or dipoles have stronger intermolecular bonds than primary bonds when the degree of polymerization, x n , exceeds about 200. Polymer molecules with weak intermolecular forces, such as dispersion forces, require longer chains, on the order of x n > 500, in order for intermolecular forces to dominate. Other than intermolecular forces, chain entanglements play an important role in determining polymer melt viscosity. As shown in Figure 4.11, there is a critical chain length, here represented by the weight-average degree of polymerization, x w , above which chain entanglements are significant enough to retard flow and increase the melt viscosity. Below this point, chain segments are too short to allow for significant entanglement effects. Typical values of the critical chain length are 610 for polyisobutylene, 730 for polystyrene, and 208 for poly(methyl methacrylate) [1]. Generally, the critical chain length is lower for polar polymers than for nonpolar polymers, again due to intermolecular forces. Intermolecular forces also play an important role in determining the compatibility of two or more polymers in a polymer blend or polymer alloy. Although the distinction between a polymer blend and a polymer alloy is still the subject of some debate, we will use the convention that a polymer alloy is a single-phase, homogeneous material (much as for a metal), whereas a blend has two or more distinct phases as a result of polymer–polymer immiscibility (cf. Section 2.3.3). In general, polymers are

Entanglement region

5.0

log h0 (N·s/m2)

b 3.0

1.0

−1.0

−3.0

tan a = 1 tan b = 3.4

a

2.0

3.0 log x¯w

4.0

Figure 4.11 Effect of polymer chain entanglements on melt viscosity. From Z. Tadmor and C. G. Gogos, Principles of Polymer Processing, Copyright  1979 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

306

TRANSPORT PROPERTIES OF MATERIALS

incompatible with each other, so that most polymer–polymer mixtures are blends, in which the primary phase is continuous, and the secondary phase exists as droplets dispersed throughout the primary phase. As the droplet sizes become smaller, the two phases become indistinguishable, and an alloy begins to form. There are some immiscible polymer blends that form bi-continuous structures as well. Needless to say, the rheological properties of polymer mixtures are complex and nearly impossible to predict. Figure 4.12 shows the viscosity of a natural rubber (NR)/poly(methyl methacrylate) (PMMA) blend (top curve) as a function of percentage NR [2]. For comparison, the predictions of four common equations are shown. The equations are as follows: ηmix = η1 φ1 + η2 φ2

(4.26)

ηmix = η2 +

(4.27)

ηmix

(4.28)

φ1 1/(η1 − η2 ) + φ2 /2η2 φ2 = η1 + 1/(η2 − η1 ) + φ1 /2η1

ln ηmix =

Shear Viscosity (MPa · s)

0.006

φ1 (α − 1 − γ φ2 ) ln η1 + αφ2 (α − 1 + γ φ1 ) ln η2 φ1 (α − 1 − γ φ2 ) + αφ2 (α − 1 + γ φ1 ) Eq. 4.26

Eq. 4.27

Eq. 4.29

Experimental

(4.29)

Eq. 4.28

0.005

0.004

0.003 0

20

40

60

80

100

Weight % of NR

Figure 4.12 Viscosity of a natural rubber (NR)/poly(methyl methacrylate) (PMMA) polymer blend and predictions of Eq. (4.26) through (4.29) at a shear rate of 333 s−1 . Adapted from Z. Oommen, S. Thomas, C. K. Premalatha, and B. Kuriakose, Polymer, 38(22), 5611–5621. Copyright  1997 by Elsevier.

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

307

where η1 and η2 are the viscosities of components 1 and 2, and φ1 and φ2 are their corresponding volume fractions. In Eq. (4.29), α and γ are adjustable parameters [3]. Other models are based upon the immiscibility of polymer blends described above, and they model the system as Newtonian drops of the dispersed polymer with concentration φ1 in a Newtonian medium of the second polymer with concentration φ2 = 1 − φ1 . There exists some concentration, φ1I = φ2I − 1, at which phase inversion takes place; that is, at sufficiently high concentration, the droplet phase suddenly becomes continuous, and the second phase forms droplets. The phase inversion concentration has been shown to correlate with the viscosity ratio, λ = η1 /η2 , and the intrinsic viscosity for at least a dozen polymer alloys and blends: λ=



φm − φ2I φm − φ1I

[η]φm

(4.30)

where φm is the maximum packing volume fraction, which is about 0.84 for most polymer alloys and blends. Much work remains to be done in this important area of blend viscosity that is a vital component of polymer processing and characterization. 4.1.4

Momentum Transport Properties of Composites

Given the vast number of possible matrix–reinforcement combinations in composites and the relative inability of current theories to describe the viscosity of even the most compositionally simple suspensions and solutions, it is fruitless to attempt to describe the momentum transport properties of composite precursors in a general manner. There are, however, two topics that can be addressed here in an introductory fashion: flow properties of matrix/reinforcement mixtures; and flow of matrix precursor materials through the reinforcement. In both cases, we will concentrate on the flow of molten polymeric materials or precursors, since the vast majority of high-performance composites are polymer-based. Furthermore, the principles here are general, and they apply to the fluid-based processing of most metal–, ceramic–, and polymer–matrix composites. 4.1.4.1 Viscosity of Polymer-Fiber Melts. For most liquids with suspended particles that are processed into reinforced matrix composites, the viscosity can be described by the modified Einstein equation given in Eq. (4.20), which accounts for the variation in particle shape using the hydrodynamic shape factor, KH . However, fiber-reinforced polymers are so ubiquitous and commercially important that some empirical studies have been performed that are worth describing here. Recent studies have included carbon [4], glass [5–7], and polyester [8] fibers in polyethylene [7], polypropylene [4, 5], polycarbonate [4, 5], polyamide [5, 6], and polyurethane [8] matrices. In addition to composition, reinforcing fibers can vary by length and diameter, but are normally classified as either long fibers (length more than about 5 mm) or short fibers (length less than about 5 mm, with a corresponding length/diameter ratio of 500–800). The effect of fiber fillers on polymer viscosity is generally the same, regardless of fiber length. As shown in Figure 4.13, the effect of the fibers is to cause more pronounced shear-thinning in the polymer (see 5 and 10 wt% curves) than is present without the fibers (0% curve), with plateaus at high and low shear rates. This is due to the tendency of fibers to align along the direction of flow, especially at high shear

308

TRANSPORT PROPERTIES OF MATERIALS

10000 T = 330 °C

h, Pa-sec

1000

100 0% 5% Long 10% Long

10 0.01

0.1

1 . Shear rate, gR (sec−1)

10

Figure 4.13 Effect of fiber content on shear viscosity for long glass fibers in polycarbonate. Reprinted, by permission, from J. P. Green and J. O. Wilkes, Polym. Eng. Sci., 32(21), 1670. Copyright  1995 Society of Plastics Engineers.

rates, thus counteracting any viscosity-building effect due to suspended particles that exists at low shear rates. The extent of shear thinning and the existence of a lowshear-rate plateau depend on the aspect ratio, fiber modulus, fiber concentration, and the viscosity of the suspending fluid. The low shear rate plateau is usually not observed for suspensions of fibers with small aspect ratios at high concentrations in lowviscosity fluids. Power-law expressions are still used to describe such polymer–fiber melts. Typical power-law parameters for selected fiber–polymer systems are shown in Table 4.7. Semiempirical expressions based on Eq. (4.23) have also been developed, as well as models based on energy dissipation. A complete review of these correlations is beyond the scope of this text, and the interested reader is referred to reference 9 for a more complete review of viscosity in fiber-reinforced polymer melts. 4.1.4.2 Viscosity of Polymers Through Fiber Preforms. The flow of molten polymers, sometimes called resins if they are network polymers, through a fibrous preform is integral to a number of important composite processes, such as melt infiltration. In addition to nonreactive processes, there are a number of important reactive processing techniques such as reaction injection molding (RIM), resin transfer molding (RTM), and reactive liquid composite molding (RLCM) in which the polymer precursors flow and polymerize in situ as they flow through the fibrous preform. These techniques are described in more detail in Chapter 7, but it is worth describing the viscosity effects here first. In the general case of a resin filling a fiber-packed mold, such as is found in melt infiltration (see Figure 4.14), fluid flow is equivalent to flow through porous

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

309

Table 4.7 Power-Law Parameters for a Fiber-Filled Polyamide Sample

K(N·sn /m2 )

n

Polyamide (PA) 6.6 PA 6.6 with 0.5 wt% blowing agent 30 vol% long glass-fiber-filled PA 40 vol% long glass-fiber-filled PA 30 vol% long glass-fiber-filled PA with 0.5 wt% blowing agent

213.1 109.2

0.81 0.80

1709.5

0.74

2400.6

0.73

1173.3

0.75

Source: S. F. Shuler, D. M. Binding, and R. B. Pipes, Polymer Composites, 15(6), (1994), 427.

P

Resin Pool L 2

X

Resin Front Fiber

Figure 4.14 Schematic diagram of the melt infiltration process for flow past a unidirectional fiber assembly.

media or packed columns, both of which have been studied for a long time. Typically, Darcy’s Law is used for flow through porous media. This model describes the flow of Newtonian fluids in porous media, and it states that the volumetric flow rate, Q, through a constant-area packed channel is proportional to the cross-sectional area, A, and the pressure drop across the channel, P , and is inversely proportional to the length of the channel, L, and the viscosity of the fluid, µ:  A P Q = KD (4.31) µ L where KD is called the permeability, and has units of length squared. Not surprisingly, Eq. (4.31) has the same form as the Hagen–Poiseuille equation for volumetric flow in a cylindrical tube, with a correction factor for the permeability of the media (KD ), which in this case is fibers.

310

TRANSPORT PROPERTIES OF MATERIALS

The permeability is given by the Kozeny–Carman relationship KD =

ε3 (1 − ε)2



1 S2k



(4.32)

where ε is the porosity, S the specific surface of the medium particle (6/D for spheres, 4/D for cylinders), D is the diameter, and k is the so-called Kozeny constant. For the specific case of flow normal to the axial direction of a parallel array of fibers with diameter, D, as is found in the melt infiltration example, the permeability becomes KD = D 2 ε3 /[80(1 − ε)2 ]. For the case of simultaneous reaction and flow, as is present in reactive molding, an expression is needed that explicitly relates the viscosity to both the temperature and the extent of reaction. The viscosity increases with time as the reaction proceeds (see Figure 4.15, top diagram). The temperature, however, has opposing effects; it will decrease the viscosity as the temperature is raised, but will also increase the reaction rate, leading to increased viscosity. To separate these effects, the polymerization reaction kinetics must be monitored independently. A viscosity–concentration correlation can be constructed by taking isochrones (t1 and t2 in Figure 4.15) of viscosity and extent of reaction and then replotting them on a viscosity–conversion diagram. The following relationship, proposed by Castro and Macosko [10], has been found useful in modeling the reactive molding process:

µ = Aµ exp(Eµ /RT )

Cg∗ Cg∗ − C ∗

A+BC ∗

(4.33)

Here Aµ and Eµ are the Arrhenius-type preexponential and activation energy terms, R is the gas constant, T is absolute temperature, Cg∗ is the solidification (gel) composition, and C ∗ is the extent of reaction, (C0 − C)/C0 .

Viscosity

Viscosity

h2 h1 Time

h2 h1

Conversion

C1*

1.0 C*2

C2*

Conversion

C*1 t1

t2

Time

Figure 4.15 Mapping of viscosity rise versus time plot into viscosity–conversion plot. Reprinted, by permission, from P. K. Mallick, Composites Engineering Handbook, p. 507. Copyright  1997 by Marcel Dekker, Inc.

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

311

Cooperative Learning Exercise 4.4

Darcy’s Law can be utilized for non-Newtonian as well as Newtonian fluids, by using an effective viscosity in place of the Newtonian viscosity: v0 =

Q KD P = A ηeff L

where v0 is the superficial velocity of the fluid flow front. In addition to the shear-rate dependent non-Newtonian viscosities described earlier in this chapter, there are other models that are stress-dependent. One such model is the Ellis Model, in which the ratio of the zero-shear-rate viscosity, η0 , to the effective viscosity, ηeff , for the case of flow through porous media is given by 4 η0 =1+ ηeff α+3



τRh τ1/2

α−1

where τ1/2 is the shear stress at which ηeff = η0 /2, τRh is the wall shear stress based on the hydrodynamic radius, and α is a constant that is the inverse of the power-law exponent, n. Person 1 : Calculate the permeability, KD , for the case of a fiber preform made from unidirectional fibers of diameter 10 µm, and a porosity of 60%. Person 2 : Derive an expression for the effective viscosity for a 1.0% solution of poly(ethylene oxide) at 293 K, which has a η0 = 2.37 N·s/m2 , τ1/2 = 0.412 N/m2 . Consult Table 4.4 for the appropriate value of α. Combine your answers to come up with an expression for the superficial velocity of a 1.0% solution of poly(ethylene oxide) within a mold consisting of 10-µm-diameter unidirectional fibers with 60% porosity. 2.37 = ; 0.88 1 + 1.8τRh P L

0.88 v0 = (7.12 × 10−13 + 1.27 × 10−12 τRh )

Answers: KD = 1.7 × 10−12 ;

ηeff

4.1.5

Momentum Transport Properties in Biologics

Many biological fluids are aqueous-based suspensions of particles. Such is the case with perhaps the most important of all biological fluids: blood. Blood is essentially a suspension of red blood cells in plasma. The concentration of red blood cells in blood is described as the hematocrit, or fraction of the blood volume represented by blood cells. The viscosity of human blood as a function of hematocrit is shown in Figure 4.16. Note that the viscosity of blood is substantially lower in small tubes than in large tubes, especially at high hematocrit values. This is due to the tendency of red blood cells to be carried in midstream (think “velocity profile” and “fully developed flow”!) such that in small blood vessels, the fluid at the wall is nearly completely plasma, which has a lower viscosity than that of whole blood (compare 0% to “normal value” hematocrit for large tubes in Figure 4.16). This has important implications to human anatomy. If the heart had to be large and powerful enough to propel blood throughout the body with “normal viscosities” such as are found in large tubes (e.g., arteries), then there would hardly be room in the chest cavity for lungs and other important organs.

312

TRANSPORT PROPERTIES OF MATERIALS

14

Viscosity (kg (m • sec)) × 103

12

Large Tubes

10 8 6 Small Tubes 4 2 Normal Value 0

0

20

40 Hematocrit (%)

60

80

Figure 4.16 Viscosity of blood as a function of percentage blood volume represented by red blood cells (hematocrit). From Biological Process Engineering, by A. T. Johnson; Copyright  1999 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Other biological fluids can be modeled as suspensions of particles in a solvent, such as was used for the description of suspensions and slurries in Section 4.1.2.2, namely, that the relative viscosity of the suspension is related to the hydrodynamic shape factor, KH , and the volume fraction of particles, fpv , as in Eq. (4.20): ηr = 1 + KH fpv

(4.20)

The biggest difference between biological particles and ceramic particles in the application of Eq. (4.20) is that while most ceramic particles are spherical (KH = 2.5), most biological particles can be modeled as either prolate ellipsoids or oblate spheroids (or ellipsoids). Ellipsoids are characterized according to their shape factor, a/b, for which a and b are the dimensions of the semimajor and semiminor axes, respectively (see Figure 4.17). In a prolate ellipsoid, a > b, whereas in an oblate ellipsoid, b > a. In the extremes, a ≫ b approximates a cylinder, and b ≫ a approximates a disk, or platelet. The hydrodynamic shape factor and axial ratio are related (see Figure 4.18), but are not generally used interchangeably in the literature. The axial ratio is used almost exclusively to characterize the shape of biological particles, so this is what we will utilize here. As the ellipsoidal particle becomes less and less spherical, the viscosity deviates further and further from the Einstein equation (see Figure 4.19). Note that in the limit of a = b, both the prolate and oblate ellipsoid give an intrinsic viscosity of 2.5, as predicted for spheres by the Einstein equation. In practice, it is the viscosity that is experimentally determined, and the correlations are used to determine axial ratios and shape factors. The viscosity can be determined by any number of techniques, the most common of which is light scattering. In addition to ellipticity, solvation (particle swelling due to water absorption) can have an effect on

MOMENTUM TRANSPORT PROPERTIES OF MATERIALS∗

313

Axis of rotation

b b

b a

Axis of rotation

b

Equatorial plane

a

Equatorial plane

(a)

(b)

Figure 4.17 (a) Prolate and (b) oblate ellipsoids of revolution, showing the relationship between the semiaxes and the axis of revolution. Reprinted, by permission, from P. Hiemenz, Principles of Colloid and Surface Chemistry, 2nd ed., p. 26. Copyright  1986 by Marcel Dekker, Inc. 4.0 Prolate

log KH

3.0 Oblate

2.0

KH = k

a b

1.81

KH = k 1.0 1.0

1.5

a b

1.00

2.0 log (a/b)

2.5

3.0

Figure 4.18 Theoretical values of the shape factor for ellipsoids. Adapted from F. H. Silver and D. L. Christiansen, Biomaterials Science and Biocompatibility, p. 150. Copyright  1999 by Springer-Verlag.

the viscosity of biological suspensions. An example of how both solvation and particle ellipticity can affect solution viscosity is shown in Figure 4.20. Finally, for biological molecules that are macromolecules, such as most proteins, Eq. (4.23) can also be used to relate the relative viscosity to the intrinsic viscosity of the solution and the macromolecule concentration: ηr = 1 + [η]c + k ′ [η]2 c2 + · · ·

(4.23)

314

TRANSPORT PROPERTIES OF MATERIALS

40

te ola

20

Pr

Intrinsic viscosity

30

ate

Obl

10

2.5 0

5

10 Axial ratio, a/b

15

20

Figure 4.19 Intrinsic viscosity as a function of axial ratio for oblate and prolate ellipsoids. Intrinsic viscosity in mL/g. Reprinted, by permission, from P. Hiemenz, Polymer Chemistry, p. 596. Copyright  1984 by Marcel Dekker, Inc. 20.0

10.0 6.0 4.0 3.0

12.0

11.0

10.0

9.0

8.0

7.0

6.0

5.0

4.5

4.0

3.5

1.0

[h] =

3.0

a b

2.0

0.6 0.4 0.3 0.2

0.1 0.06 0.4

0.8

1.2

1.6 grams water gram protein

2.0

2.4

2.8

Figure 4.20 Variation of intrinsic viscosity of aqueous protein solutions with axial ratio and extent of solvation. Reprinted, by permission, from P. Hiemenz, Polymer Chemistry, p. 598. Copyright  1984 by Marcel Dekker, Inc.

HEAT TRANSPORT PROPERTIES OF MATERIALS

315

Cooperative Learning Exercise 4.5

Some plots of reduced specific viscosity (see Table 4.5) as a function of concentration for various samples of vitreous hyaluronan bovine protein are shown at right. Hyaluronan may be modeled as a prolate ellipsoid, but it is unknown whether it is a flexible or rigid chain. 500

400

hsr (ml/g)

Person 1 : Pick one of the curves in the plot, and determine the intrinsic viscosity by applying Eq. (4.23) and the definition of reduced specific viscosity. Use this value of intrinsic viscosity to estimate the axial ratio of the particle using Figure 4.19. Note that you may have to assume a linear approximation to prolate ellipsoid curve and extrapolate to obtain an estimate for a/b. Person 2 : Use the a/b ratio estimated by Person 1 to estimate the diffusivity of this protein. The diffusivity of prolate ellipsoids, DAB , in units of cm2 /s, can be estimated using the following relationship

300

200

100

0

DAB

(2.15 × 10−13 cm3 /s)[ln(2a/b)] = a

1

3 2 4 Concentration (mg/ml)

Reprinted, by permission, from F. H. Silver and D. L. Christiansen, Biomaterials Science and Biocompatibility, p. 151. Copyrigyt © 1999 by Springer-Veralg.

For ease of calculation, assume that the hyaluronan protein has a chain length, a = 215 nm. Compare the intrinsic viscosity and diffusivity results to come up with some conclusion regarding the shape and rigidity of this protein; i.e., rigid chain or flexible chain. Answers: The intrinsic viscosity of all the plots are determined by extrapolating back to zero concentration, and are found to range from about 110 to 410 mL/g. According to Figure 4.19, this corresponds to a/b values on the order of 1000 or so. When this value is plugged into the equation for diffusivity, one obtains a value of about 10−12 m2 /s. As described in the text, the intrinsic viscosity is relatively high, and the diffusivity is relatively low, leading us to believe that the hyaluronan chain is somewhat rigid.

An example of how both particle concentration and axial ratio can be used to describe their effect on solution viscosity—in this case the reduced viscosity—is shown in Figure 4.21. Here, the deviations from the Einstein model (a/b = 1 at low concentrations) are high, since it is based upon a dilute solution of nonsolvated, spherical particles. Later in the chapter, we will see that there is a relationship between particle shape and size, intrinsic viscosity, and the ability of the particle to diffuse through the solution, as characterized by the diffusion coefficient, or diffusivity. In general, molecules with high values of intrinsic viscosity (>100 ml/g) and relatively low diffusivities (10−9 m2 /s) and low intrinsic viscosities are roughly spherical. Consult Section 4.3.5 for more information on diffusivities of biologics.

316

TRANSPORT PROPERTIES OF MATERIALS

1.20

a = 64.3 b

a = 31.5 b

hr

1.15

a = 20.3 b

1.10 1.05 1.00

a =1 b 1.0

2.0

3.0

4.0

5.0

fpv × 103

Figure 4.21 Effect of axial ratio and particle concentration on relative viscosity. Data are for tobacco mosaic virus particles. Adapted from M. A. Lauffer, J. Am. Chem. Soc., 66, 1188. Copyright  1944 by The American Chemical Society, Inc.

4.2

HEAT TRANSPORT PROPERTIES OF MATERIALS

Of the three general categories of transport processes, heat transport gets the most attention for several reasons. First, unlike momentum transfer, it occurs in both the liquid and solid states of a material. Second, it is important not only in the processing and production of materials, but in their application and use. Ultimately, the thermal properties of a material may be the most influential design parameters in selecting a material for a specific application. In the description of heat transport properties, let us limit ourselves to conduction as the primary means of transfer, while recognizing that for some processes, convection or radiation may play a more important role. Finally, we will limit the discussion here to theoretical and empirical correlations and trends in heat transport properties. Tabulated values of thermal conductivities for a variety of materials can be found in Appendix 5. 4.2.1

Heat Transport Properties of Metals and Alloys

4.2.1.1 Conduction, Convection, and Radiation. We begin this section with a brief review of the three primary modes of heat transfer: conduction, convection, and radiation. In conduction, heat is conducted by the transfer of energy of motion between adjacent molecules in a liquid, gas, or solid. In a gas, atoms transfer energy to one another through molecular collisions. In metallic solids, the process of energy transfer via “free electrons” is also important. In convection, heat is transferred by bulk transport and mixing of macroscopic fluid elements. Recall that there can be forced convection, where the fluid is forced to flow via mechanical means, or natural (free) convection, where density differences cause fluid elements to flow. Since convection is found only in fluids, we will deal with it on only a limited basis. Radiation differs from conduction and convection in that no medium is needed for its propagation. As a result, the form of Eq. (4.1) is inappropriate for describing radiative heat transfer. Radiation is

HEAT TRANSPORT PROPERTIES OF MATERIALS

317

the transfer of energy through space by means of electromagnetic waves. The laws that govern the movement and absorption of electromagnetic waves such as light govern the transfer of heat by radiation. We will cover some of these topics later in Chapter 6. 4.2.1.2 The Molecular Origins of Thermal Conductivity. In an analogous fashion to the derivation from the potential energy function for the viscosity of a dilute gas at low density [see Eq. (4.6], the thermal conductivity of a gas, k, based on the rigid sphere model, can be derived for a gas (in W/m·K): −2

k = 8.32 × 10



T /M σ 2

(4.34)

where T is the absolute temperature in degrees kelvin, M is the atomic weight of the monatomic gas, and σ and  are the same Lennard-Jones parameters used in Eq. (4.6). Notice the dependence on the same properties as viscosity—namely, atomic weight, molecular diameter, and temperature—although the thermal conductivity has an inverse dependence on the square root of molecular weight instead of the proportional dependence for viscosity. There is no variation of thermal conductivity with pressure for a rigid sphere, ideal, monatomic gas at low pressure, but the model does predict that thermal conductivity will be zero at absolute zero. It can be shown that thermal conductivity and viscosity are related in this case as follows: 5Cˆ v µ 15Rµ = (4.35) k= 4M 2 where M is again the atomic weight of the gas, R is the gas constant in appropriate units [1.9872 cal/mol·K for thermal conductivity (in cal/cm·K·s) and 831.4 J·cm/m·mol·K for thermal conductivity (in W/m·K)], and µ is the viscosity of the gas in poise (g/cm·s). The quantity, Cˆ v , is called the heat capacity per unit mass at constant volume and has the same units as R/M. In the final section of this chapter, we will see how both the viscosity and thermal conductivity (thermal diffusivity) are related to the molar diffusivity. As with viscosity, the theoretical predictions become more complex as the atoms themselves become more complex and more dense. For a polyatomic gas, the thermal conductivity is given by extension of Eq. (4.35): k=

15Rµ 4M



3 4Cv + 15R 5



(4.36)

where R, µ, and M are as in Eq. (4.35) (with M the molecular weight of the polyatomic gas rather than the atomic weight). Here, Cv is the molar heat capacity, obtained by multiplying the per mass heat capacity by the molecular weight, M Cˆ v . Note that for an ideal monatomic gas, Cv = 3R/2, and Eq. (4.36) reduces to (4.35). Equation (4.36) provides a simple method for estimating an important heat transfer dimensionless group called the Prandtl number. Recall from general chemistry and thermodynamics that there are two types of molar heat capacities, Cv , and the constant pressure heat capacity, Cp . For an ideal gas, Cv = 3Cp /5. The Prandtl number is

318

TRANSPORT PROPERTIES OF MATERIALS

defined as Pr = Cˆ p µ/k. Substitution of Eq. (4.36) into this relation, along with use of the relationship between Cv and Cp , gives the following for polyatomic gases: Pr =

Cp Cp + 1.25R

(4.37)

Cooperative Learning Exercise 4.6

The molecular weight of O2 is 32.000; its molar heat capacity Cp at 300 K and low pressure ˚ is 29.37 J/mol·K. The Lennard-Jones parameters for molecular oxygen are σ = 3.433 A and  = 1.074. Person 1 : Estimate the viscosity of molecular oxygen at low pressure and 300 K. Then estimate the Prandtl number, Pr, using Eq. (4.37). Person 2 : Use Person 1’s estimate of viscosity to estimate the thermal conductivity of molecular oxygen at low pressure and 300 K using Eq. (4.36). Note that you will need to modify Eq. (4.36) to include Cp instead of Cv . Beware of units and molar heat capacities! Combine your answers to give an estimate for the Prandtl number, Pr, based on the definition Pr = Cˆ p µ/k. How does this compare with your estimation given by Eq. (4.37)? Answers: Person 1: µ = 2.066 × 10−4 g/cm·s; Pr = 0.74. Person 2: k = 0.023 W/m·K. Combined: Pr = 0.66.

The introduction of heat capacity into the relationships for thermal conductivity and the Prandtl number gives us an opportunity to make a clarification regarding these two quantities. Thermal conductivity is a true heat transport property; it describes the ability of a material to transport heat via conduction. Heat capacity, on the other hand, is a thermodynamic quantity and describes the ability of a material to store heat as energy. The latter, while not technically a transport property, will nonetheless be described in this chapter for the various materials types, due in part to its theoretical relationship to thermal conductivity, as given by Eq. (4.35) and (4.36), and, more practically, because it is often used in combination with thermal conductivity as a design parameter in materials selection. 4.2.1.3 Thermal Conductivities of Liquids. As was the case with viscosity, it is difficult to derive useful relationships that allow us to estimate thermal conductivities for liquids from molecular parameters. There is a theoretical development by Bridgman, the details of which are presented elsewhere [11], which assumes that the liquid molecules are arranged in a cubic lattice, in which energy is transferred from one lattice plane to the next at sonic velocity, vs . This development is a reinterpretation of the kinetic theory model used in the last section, and with some minor modifications to improve the fit with experimental data, the following equation results:

k = 2.8(V /N )−2/3 kB vs

(4.38)

HEAT TRANSPORT PROPERTIES OF MATERIALS

319

where V /N is the volume per molecule and kB is Boltzmann’s constant. This equation is limited to liquids well above their critical densities. The sonic velocity in liquids can be estimated by   Cp ∂p vs = (4.39) Cv ∂ρ T where the quantity (∂p/∂)T may be obtained from isothermal compressibility measurements or from an equation of state, and (Cp /Cv ) is approximately unity for liquids away from their critical point. More frequently, empirical relationships are used to estimate liquid thermal conductivities. A linear relationship with temperature is adequate, since liquid thermal conductivities do not vary considerably with temperature, and hardly at all with pressure. Some experimental data for thermal conductivities of pure liquid metals are shown in Figure 4.22, and data for some binary metallic alloys are shown in Figure 4.23. These values will, in general, be lower than the thermal conductivities of the corresponding solid forms of the metals and alloys. 4.2.1.4 Thermal Properties of Metallic Solids. In the preceding sections, we saw that thermal conductivities of gases, and to some extent liquids, could be related to viscosity and heat capacity. For a solid material such as an elemental metal, the link between thermal conductivity and viscosity loses its validity, since we do not normally think in terms of “solid viscosities.” The connection with heat capacity is still there, however. In fact, a theoretical description of thermal conductivity in solids is derived directly from the kinetic gas theory used to develop expressions in Section 4.2.1.2.

180 Cu

Thermal conductivity/W m−1 K−1

160 140 120

Al

100 80 60 40

Sn

20 0

Na

Ga

Zn Te Li

Sb

Pb Bi

Hg 0

K

2

4

6

8

10

12

14

16

18

20

2 °C

Temperature/10

Figure 4.22 Thermal conductivity of some pure liquid metals as a function of temperature. Reprinted, by permission, from T. Iida and R. I. L. Guthrie, The Physical Properties of Liquid Metals, p. 241. Copyright  1988 by Oxford University Press.

TRANSPORT PROPERTIES OF MATERIALS

Thermal conductivity/W m−1 K−1

320

Sn−Pb (38 wt% Pb)

30

K−Na (43.5 wt% Na) K−Na. (23 wt% Na)

Na−Hg (6.3 wt% Hg)

20

Bi−Pb eutectic (44.5 wt% Pb) 10 Na−Hg (30 wt% Hg) 0

0

200

400

600

800

Temperature/°C

Figure 4.23 Thermal conductivity of some liquid alloys as a function of temperature. Reprinted, by permission, from T. Iida, and R. I. L. Guthrie, The Physical Properties of Liquid Metals, p. 241. Copyright  1988 by Oxford University Press.

Equation (4.34), for example, can be derived from a simple expression for the thermal conductivity of a monatomic gas: k=

1 ˆ ρ Cv vl 3

(4.40)

where ρ is the gas density, Cˆ v is the constant volume heat capacity per unit mass (sometimes called the specific heat), v is the mean molecular speed (sometimes called the average particle velocity), and l is a quantity called the mean free path, which is a measure of the average distance traveled by the molecules between collisions. We saw in Section 4.2.1.2 that the molar heat capacity, Cv , equals 3R/2 for an ideal monatomic gas. This is because almost all of the energy of the gas is kinetic in nature; that is, an increase in temperature results in a direct increase in the translational kinetic energy of the atoms. For polyatomic gases, additional energy states are available, (e.g., vibrational and rotational, such that the gas can store energy in forms other than translational motion), and a higher heat capacity than that predicted by 3R/2 results. In a solid, the atoms are similar to a vibrating diatomic molecule, where half of the thermal energy is the kinetic energy (3R/2) and half is the average potential energy (3R/2). In this simplistic approach, the heat capacity should then be roughly 3R for solids. This approximation is known as the Law of Dulong–Petit, and it is applicable to many nonmetallic solids at and above room temperature. It does not apply at low temperatures because the energy of atomic vibrations is quantized and the number of vibrational modes that are excited decreases as the temperature decreases below some value for each material. This temperature is known as the Debye temperature, θD , above which the heat capacity approaches the value of Dulong and Petit (3R), and below which the heat capacity decreases monotonically with decreasing temperature

HEAT TRANSPORT PROPERTIES OF MATERIALS

321

3R

Cv

2R

R

0

0

2qD

qD

T

Figure 4.24 Molar heat capacity as a function of temperature, based on the Debye model. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

Table 4.8 Representative Values of Debye Temperature for Selected Elements Element

θD (K) Element θD (K)

Ag 225 Al 418 Au 165 Be 1160 Bi 117 C (diamond) 2000 Ca 219 Cd 300 Co 445 Cr 402 Cu 339

Fe Ge Hg In La Mg Mo Nb Ni Pb Pd

467 366 60–90 109 132 406 425 252 456 94.5 275

Element

θD (K)

Pt Si Sn (gray) Sn (white) Ta Ti Tl V W Zn Zr

229 658 212 189 231 278 89 273 379 308 270

Source: C. Kittel, Introduction to Solid State Physics, p. 132, 2nd ed. Copyright  1957 by John Wiley & Sons, Inc.

(see Figure 4.24 and Table 4.8 for values of θD ). In this region (T ≪ θD ), the heat capacity is due to lattice vibrations (hence, a subscript l is added), and is given by: 12π 4 R Cvl = 5



T θD

3

(4.41)

Most nonmetals follow Eq. (4.41) at low temperatures, as illustrated in Figure 4.24. Most metals, however, show a linear decrease in heat capacity with temperature as

322

TRANSPORT PROPERTIES OF MATERIALS

absolute zero is approached. We require a slightly different model to describe the lowtemperature behavior of Cv in metals. Recall from Figure 1.15 that metals have free electrons in what is called the valence band and have empty orbitals forming what is called the conduction band. In Chapter 6, we will see how this electronic structure contributes to the electrical conductivity of a metallic material. It turns out that these same electronic configurations can be responsible for thermal as well as electrical conduction. When electrons act as the thermal energy carriers, they contribute an electronic heat capacity, Cve , that is proportional to both the number of valence electrons per unit volume, n, and the absolute temperature, T : π 2 nkB2 T (4.42) Cve = 2EF where kB is Boltzmann’s constant and EF is the energy of the Fermi level, which is characteristic of an element, but in general is between 3 and 5 eV. The contribution of the electronic heat capacity to the overall heat capacity is usually quite small, but below about 4 K the electronic contribution becomes larger than the lattice contribution [Eq. (4.41)]. In this low-temperature regime, the total heat capacity is a sum of both the electronic and lattice contributions: Cv = Cvl + Cve = αT 3 + γ T

(4.43)

where α is given by the coefficients in Eq. (4.41), and γ is given by the coefficients in Eq. (4.42). Returning now to thermal conductivity, Eq. (4.40) tells us that any functional dependence of heat capacity on temperature should be implicit in the thermal conductivity, since thermal conductivity is proportional to heat capacity. For example, at low temperatures, we would expect thermal conductivity to follow Eq. (4.43). This is indeed the case, as illustrated in Figure 4.25. In copper, a pure metal, electrons are the primary heat carriers, and we would expect the electronic contribution to heat capacity to dominate the thermal conductivity. This is the case, with the thermal conductivity varying proportionally with temperature, as given by Eq. (4.42). For a semiconductor such as germanium, there are less free electrons to conduct heat, and lattice conduction dominates—hence the T 3 dependence on thermal conductivity as suggested by Eq. (4.41). This competition between electrons and the heat carriers in the lattice (phonons) is the key factor in determining not only whether a material is a good heat conductor or not, but also the temperature dependence of thermal conductivity. In fact, Eq. (4.40) can be written for either thermal conduction via electrons, ke , or thermal conduction via phonons, kp , where the mean free path corresponds to either electrons or phonons, respectively. For pure metals, ke /kp ≈ 30, so that electronic conduction dominates. This is because the mean free path for electrons is 10 to 100 times higher than that of phonons, which more than compensates for the fact that Cve is only 10% of the total heat capacity at normal temperatures. In disordered metallic mixtures, such as alloys, the disorder limits the mean free path of both the electrons and the phonons, such that the two modes of thermal conductivity are more similar, and ke /kp ≈ 3. Similarly, in semiconductors, the density of free electrons is so low that heat transport by phonon conduction dominates.

HEAT TRANSPORT PROPERTIES OF MATERIALS

323

k ∼T

k (W/m • K)

10,000

Cu

1000

k ∼T 3 Ge

100

10

1

10

100

T |K|

Figure 4.25 Temperature dependence of thermal conductivity for a pure metal (Cu) and a nonmetal (Ge), illustrating different temperature dependences at low temperature. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

So, in general, highly ordered materials such as pure metals, most pure ceramics, and crystalline polymers exhibit increasing thermal conductivity with decreasing temperature due to increases in both the electronic and phonon mean free paths at low temperatures. At sufficiently low temperatures, the mean free paths become constant, and the variation in heat capacity, as given by Eqs. (4.41) and (4.42), dominate the thermal conductivity. Disordered materials, on the other hand, such as solid–solution alloys, inorganic glasses, and glassy polymers, exhibit decreasing thermal conductivity with decreasing temperature. This is because the electronic and phonon mean free paths are relatively small at all temperatures, and temperature variation in heat capacity dominates the temperature dependence of thermal conductivity (see Figure 4.24). These trends are summarized in Figure 4.26. Note that highly ordered materials such as copper, α-iron, polyethylene, and BeO have significant regions in which thermal conductivity decreases with temperature. Disordered materials such as PMMA, glass, and stainless steel tend to increase thermal conductivity with temperature. The lowering of thermal conductivity with alloying is an important design consideration, as illustrated in Figure 4.27. Note that as more alloying elements are added to iron, the thermal

324

TRANSPORT PROPERTIES OF MATERIALS

103 Cu

k (W/m • K)

102

Dense

a-Fe

BeO

Type 304 stainless steel 101 Silica glass 100 PE PMMA −1

10

0

200

400

600

800

1000

T |K|

Figure 4.26 Variation of thermal conductivity with temperature for ordered (Cu, BeO, Fe, and PE) and disordered (SS, silica glass, and PMMA) with temperature. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

conductivity drops at a given temperature. In the subsequent sections, we expand upon these generalizations for the remaining material classes. 4.2.2

Heat Transport Properties of Ceramics and Glasses

4.2.2.1 Heat Transport Properties of Ceramics. As a means of further illustrating the importance of the theoretical developments for heat capacity and thermal conductivity in the previous section, Figure 4.28 shows that the general trend of increasing heat capacity with temperature and the ultimate achievement of Cv = 3R shown in Figure 4.24 holds true for crystalline ceramic materials as well. The primary result of heat capacity theory as applied to ceramics is that the 3R value is achieved in the neighborhood of 1000◦ C for most oxides and carbides. Further increases in temperature do not strongly affect this value, and it does not depend strongly on the crystal structure. However, heat capacity does depend on porosity. Consequently, the heat energy required to raise the temperature of porous insulating fire brick is much lower than that required to raise the temperature of the dense fire brick material, even though its thermal conductivity is lower. This is one of the useful properties of insulating materials for furnaces, which can be rapidly heated and cooled due to their porosity. Other factors affecting heat capacity in ceramics are phase transformations, such as the α –β quartz transition, which result in step changes to the heat capacity, and the

HEAT TRANSPORT PROPERTIES OF MATERIALS

325

THERMAL CONDUCTIVITY 1000 800

Iron alloys

600 400 Thermal conductivity (milliwatts/cm deg K)

0.1%

Mild

C 2% Ni

200

3%

Ni

2%

Mn

13%

100

57%

Ni

80 60 40

Ni % 31 Ni % 36

20

%

27

10 10

20

%

13

or

,q Cr

%

19 %

13

, Mn

d

he

nc

ue

Mn

39%

13% Cr 4% Al

Mn

Cr 1%

%

24

Cr

Ni

or

Ni

40 60 Temperature (°K)

80

100

200

Figure 4.27 Thermal conductivity of some iron alloys. Reprinted, by permission, from C. Kittel, Introduction to Solid State Physics, 2nd ed., p. 151. Copyright  1957 by John Wiley & Sons, Inc.

Heat capacity (cal/g atom °C)

9 8 7 6

SiC

MgO 3R

5

Al2O3

Mullite

4 3 2 1 0

−200 −100 0

100 200 300 400 500 600 700 800 900 10001100120013001400 Temperature (°C)

Figure 4.28 Heat capacity of some ceramic materials at different temperatures. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

326

TRANSPORT PROPERTIES OF MATERIALS

presence of point defects (cf. Section 1.1.3), which tend to increase the value of the heat capacity. The discussion of the previous section would also lead us to believe that since most ceramics are poor electrical conductors (with a few notable exceptions) due to a lack of free electrons, electronic conduction would be negligible compared to lattice, or phonon, conduction. This is indeed the case, and we will see that structural effects such as complexity, defects, and impurity atoms have a profound effect on thermal conductivity due to phonon mean free path, even if heat capacity is relatively unchanged. Materials with complex structures possess more thermal phonon scattering, which reduces the mean free path and lowers the thermal conductivity, as given by Eq. (4.40). For example, the magnesium aluminate spinel shown in Figure 1.41 has a lower thermal conductivity than either Al2 O3 or MgO, although each has a similar heat capacity and structure. Similarly, the complex structure of mullite has a much lower thermal conductivity than magnesium aluminate spinel. Structure can also affect the temperature dependence of thermal conductivity, since in complex structures the phonon mean free path tends to approach lattice dimensions at high temperatures. Similarly, anisotropic crystal structures can have thermal conductivities that vary with direction, as illustrated for quartz (cf. Figure 1.44) in Table 4.9. This variation tends to decrease with increasing temperature, since anisotropic crystals become more symmetrical as the temperature is raised. Crystallite size also has an effect on thermal conductivity, though its effect is evident only at higher temperatures due to electronic conductivity, as shown in Figure 4.29 for three common ceramic materials. Finally, in the same way that structural complexity lowers conductivity, the addition of impurities and variations in composition can affect thermal conductivity. As shown in Figure 4.30, the thermal conductivity is a maximum for simple elementary oxide and carbide structures, such as BeO. As the cation atomic number increases and the difference in atomic weight between the components becomes larger, the thermal conductivity decreases. In solid solutions, the additional phonon scattering for low concentrations of impurities is directly proportional to the volume fraction of impurity, and the impurity effect is relatively independent of temperature. The effect of solid–solution impurities over an extended range of compositions depends on mass difference, size difference, and binding energy difference of the constituents. In addition, temperature effects can be magnified in solid solutions, as illustrated in Figure 4.31. For the Table 4.9 Thermal Conductivity of Quartz Thermal Conductivity (W/m·K) Temperature (◦ C)

Normal to c-Axis

Parallel to c-Axis

Ratio

0 100 200 300 400

0.67 0.50 0.42 0.35 0.31

1.13 0.79 0.63 0.50 0.42

1.69 1.58 1.50 1.43 1.35

Source: W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc.

HEAT TRANSPORT PROPERTIES OF MATERIALS

327

Thermal conductivity (cal sec−1 cm ° C−1 cm−2)

0.04

0.03 Single-crystal TiO2 II to c axis

Single-crystal Al2O3

Polycrystalline Al2O3 17m grain size 9m grain size

0.02

Single-crystal TiO2 ⊥ to c axis

0.01

Polycrystalline TiO2 Single-crystal CaF2 Polycrystalline CaF2 28m 0

0

200

400

600 800 Temperature (° C)

1000

1200

Thermal conductivity (Btu/sec/° F/ft2/in.)

Figure 4.29 Thermal conductivity of single-crystal and polycrystalline Al2 O3 , TiO2 , and CaF. Multiply by 418.7 to obtain k in units of W/m·K. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc. 0.30 C 0.10 Si

Be

Carbides at 1100° F

B 0.03 Mg Al

0.01

0.003

Oxides at 1475 ° F 5

10

Ti Ca Ti

Zn Ni

Th U

30 100 300 Atomic weight of cation

Figure 4.30 Effect of cation atomic number on thermal conductivity of some oxides and carbides. Multiply by 519.2 to obtain k in units of W/m·K. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

328

TRANSPORT PROPERTIES OF MATERIALS

Thermal conductivity (Cal /Sec /cm2/ °C/ cm)

0.07

0.06

0.05

0.04

0.03

0.02 200 °C 0.01 1000 °C 0 MgO

20

40

60

Volume percent NiO

80

100 NiO

Figure 4.31 Thermal conductivity in the solid solution system MgO–NiO. Multiply by 418.7 to obtain k in units of W/m·K. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

MgO–NiO solid solution shown, the effect of thermal conductivity lowering is even more pronounced at 200◦ C than at 1000◦ C. One final note is appropriate for this section. Due to the fact that many oxide ceramics are used as insulating materials, the term thermal resistivity is often used instead of thermal conductivity. As will be the case with electrical properties in Chapter 6, resistivity and conductivity are merely inverses of one another, and the appropriateness of one or the other is determined by the context in which it is used. Similarly, thermal conductance is often used to describe the thermal conductivity of materials with standard thicknesses (e.g., building materials). Thermal conductance is the thermal conductivity divided by the thickness (C = k/L), and thermal resistance is the inverse of the product of thermal conductance and area (R = 1/C·A). 4.2.2.2 Heat Transport Properties of Glasses. As was described in Chapter 1, glasses are structurally different from crystalline ceramics, and we would expect this to affect the heat capacity. This is true, and in general the heat capacities of most oxide glasses approach 0.7 to 0.95 of the theoretical 3R value at the low-temperature end of the glass transition. On passing through the glass transition, the heat capacity generally increases by a factor of 1.3 to 3.0, as a result of increased configurational entropy. Eventually, the heat capacity of glasses reaches a value of n3R, where n is the number of atoms in the compound; thus n = 3 for SiO2 and Cp ≈ Cv ≈ 78 J/mol·K above 1000◦ C. A number of empirical correlations exist for estimating the heat capacity of glasses and related thermodynamic quantities such as enthalpy. One such correlation is

329

HEAT TRANSPORT PROPERTIES OF MATERIALS

presented here, and others can be found in reference 14. The molar heat capacity of a multicomponent glass up to about 900◦ C can be estimated from the molar composition, xi (entered in mol%), according to

Cp =



1  100



ai xi + T



bi xi +



ci xi

T2

+



di xi



(4.44)



T 1/2

where T is the temperature in K, and the factors ai to di are presented in Table 4.10. Based on the discussion of the previous section, we would also expect that structural complexity and disorder present in glasses would tend to decrease the thermal conductivity. This is true, especially when alkalis are added to sodium silicate glasses. Figure 4.32 shows the effect of replacing SiO2 in Na2 O–SiO2 glass with other oxides. Table 4.10 Factors Used to Determine Molar Heat Capacities of Glasses Using Eq. (4.44)

Oxide

ai

103 bi

10−5 ci

di

Validity Region (K)

Na2 O K2 O MgO CaO Al2 O3 SiO2 TiO2 FeO Fe2 O3

70.884 84.323 46.704 39.159 175.491 127.200 64.111 31.770 135.250

26.110 0.731 11.220 18.650 −5.839 −10.777 22.590 38.515 12.311

−3.5820 −8.2980 −13.280 −1.5230 −13.470 4.3127 −23.020 −0.012 −39.098

0 0 0 0 −1370.0 −1463.9 0 0 0

270–1170 270–1190 270–1080 270–1130 270–1190 270–1600 300–800 300–800 300–800

W/(mK)

Source: H. Scholze, Glass. Copyright  1991 by Springer-Verlag.

1.0

B2O3

MgO

Al2O3

ZnO, PbO

k

0.9 Na2O

CaO

0.8

0

10

20

Wt %

30 0

Rm O n

10

20

Wt %

30

Rm O n

(a)

(b) ◦

Figure 4.32 Change in thermal conductivity at 0 C of an Na2 O–SiO2 glass (18–82 wt%) as SiO2 is replaced proportionally by weight with other oxides. Reprinted, by permission, from H. Scholze, Glass, p. 363. Copyright  1991 by Springer-Verlag.

330

TRANSPORT PROPERTIES OF MATERIALS

These results reflect the fact that the stronger the bonds in the glass are, the higher the thermal conductivity. The dependence of thermal conductivity on bond strength allows an empirical expression to be developed in which the thermal conductivities of glass components are additive:  ki xi (4.45) k = 10−1 where xi is the weight fraction of component i, and ki is its corresponding thermal conductivity factor, representative values for which can be found in Table 4.11. An alternative set of factors at higher temperatures has been developed by Primenko [13] and can be found elsewhere [14]. As a rough approximation, thermal conductivity of a glass can be calculated from the density, ρ, by using k=

a +b ρ

(4.46)

where the pair values for constants a and b of 2.09 and 0.17, or 2.30 and 0.21, are valid for temperatures between 0◦ C and 100◦ C. 4.2.3

Heat Transport Properties of Polymers

As with other disordered materials, the thermal conductivities of polymers are low due to phonon scattering. As a result, even though polymers tend to have heat capacities of the same order of magnitude as metals (1.5 to 3.5 J/g·K), their thermal conductivities (0.1 to 1.0 W/m·K) are 1000 times lower than metals. Polymers, therefore, are generally good insulators, as long as their use temperature is below their thermal stability temperature. Few correlations for heat capacity and thermal conductivity of polymers Table 4.11 Factors Used to Determine Thermal Conductivity of Glasses Based on Composition Using Eq. (4.45) Oxide

ki (W/m·K)a

Li2 O Na2 O K2 O MgO CaO SrO BaO B2 O3 Al2 O3 SiO2 TiO2 ZnO PbO

−9.29 −4.79 2.17 21.73 13.06 8.63 2.89 8.216 13.61 13.33 −31.38 8.00 2.68

a

Values from Ammar et al. [12]. Source: H. Scholze, Glass. Copyright  1991 by SpringerVerlag.

HEAT TRANSPORT PROPERTIES OF MATERIALS

331

Cooperative Learning Exercise 4.7

Consider a glass with the following composition (in wt%): 72 SiO2 , 15 Na2 O, 11 CaO, and 2 Al2 O3 , and a density of 2.458 g/cm3 . Person 1: Estimate the thermal conductivity at 30◦ C using Eq. (4.45). Person 2 : Estimate the thermal conductivity at 30◦ C using Eq. (4.46). Compare your answers. What are some ways you could experimentally verify these estimates? Answers: k1 = 1.06; k2 = 1.02 Thermal conductivities can be determined using any setup that would allow Fourier’s Law of conduction to be applied. For example, a plate of the glass could be placed on top of a heat sink (such as a large copper block), heated on one side, and the temperature difference at each face measured to determine the temperature gradient. Once the heat flow is measured, the proportionality constant, k, can be calculated.

exist. As for all the materials in this chapter, representative values of thermal conductivity can be found in Appendix 5. The heat capacity of polymers at constant pressure increases with temperature up to the melting point. In practice, there is a sharp increase in heat capacity at the onset of melting, followed by a sharp drop at the end of the melting range (see Figure 4.33). There is then a slight increase in heat capacity with temperature in the melt. The thermal conductivity of polymers is molecular-weight-dependent, and it increases with increasing molecular weight. Orientation also has an effect, and it can have a strong influence on thermal conductivity in highly crystalline polymers such as HDPE. As is evident from Figure 4.34, the thermal conductivity of most polymeric materials is, with a few notable exceptions for highly crystalline polymers such as polyethylene, less dependent upon temperature and, as a result, can only be influenced significantly through the incorporation of fillers. These effects will be discussed in the next section. Foaming with air or some other gas can also decrease thermal conductivity (see Figure 4.35). This

Specific heat (cal/g⋅ °C)

1.3 Dalvor 8200

1.1 0.9

Commercial PVF2

0.7

Tm

0.5 0.3

100

150

190

Temperature (°C)

Figure 4.33 Specific heat of two commercial polyvinylidene fluoride samples with temperature. From Z. Tadmor and C. G. Gogos, Principles of Polymer Processing, Copyright  1979 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

332

TRANSPORT PROPERTIES OF MATERIALS

0.50 0.55 0.50

0.40

0.45

POM

0.35

0.40

Low-

0.30

density PE

0.25

High-density PE

0.35 0.30

PC

0.25

0.20

PMMA

PP

0.15

0.20

0.10

0.15 0.116 0.10

0.05

0.05

PVC and PS

0

50

100

150

Thermal conductivity (W/m K)

Thermal conductivity (Kcal/m h °C)

0.45

0 200

Temperature (°C)

Figure 4.34 Temperature dependence of thermal conductivity for selected polymers. Reprinted, by permission, from P. C. Powell and A. J. I. Housz, Engineering with Polymers, p. 288. Copyright  1998 by P. C. Powell and A. J. I. Housz.

Thermal conductivity, W/m°C at 21°C

0.042

0.040

0.038

0.036

0.034

0

50

100

150

Density, kg/m3

Figure 4.35 Thermal conductivity of expanded polystyrene beads as a function of density. Reprinted, by permission, from The Dow Chemical Company. Copyright  1966.

333

HEAT TRANSPORT PROPERTIES OF MATERIALS

has the same effect as the introduction of porosity in ceramic materials such as fire brick, due to the introduction of a lower thermal conductivity material (in this case, a gas) into the structure.

Cooperative Learning Exercise 4.8

Recall from the beginning of the chapter that a related quantity to the thermal conductivity is the thermal diffusivity, α, which is defined as k/ρCp , where k is the thermal conductivity, ρ is the density and Cp is the heat capacity at constant pressure per unit mass, or specific heat. Below are these thermal properties for polycarbonate.

1.20

r

5.4 5.0

0.4

k

4.4

Cp

0.3

4.0 350

300

400

Cp (cal/g · °K)

1.10

k × 104 (cal/cm · s · °K)

3

1.15

14

a

10

a × 104 (cm2/s)

16 0.5

r (g/cm )

Person 1 : Determine the density and thermal diffusivity for polycarbonate at 350◦ C from the plots. Person 2 : Determine the thermal conductivity and heat capacity for polycarbonate at 350◦ C from the plots. Combine your information to calculate the thermal diffusivity at 350◦ C. How does this compare to the value obtained from the plot? Can you estimate the melting point of this polycarbonate from these plots?

8 6

450

Temperature (°K)

Reprinted, by permission, from Z. Tadmor and C. G. Gogos. Principles of Polymer Processing, p. 132. Copyright © 1979 by John Wiley & Sons, Inc.

Answers: ρ = 1.19 g/cm3 ; α = 12 × 10−4 cm2 /s; k = 4.8 × 10−4 cal/m·s·K; Cp = 0.33 cal/g·K αcalc = k/ρCp = 4.8 × 10−4 cal/m·s·K/(1.19 g/cm3 × 0.33 cal/g·K) = 12 × 10−4 cm2 /s, same as on graph. Tm ≈ 434 K = 162◦ C, as determined from graphs where discontinuity exists; thermal properties change!

4.2.4

Heat Transport Properties of Composites

As was the case with momentum transport, the heat transport properties of heterogeneous systems are difficult to correlate and virtually impossible to predict. There are two topics worthy of note, however, namely, the heat transport properties of filled composites, and the thermal conductivity of laminar composites. 4.2.4.1 Thermal Conductivity of Filled Composites. As we saw at the end of the previous section, the presence of gases in a material’s structure can dramatically affect the heat transport properties. We expand upon this principle here, again concentrating on polymeric materials that have been reinforced with various fibers and fillers and that may also contain porous gas structures, as is found in foams. To a first approximation, the heat capacity in these heterogeneous, multiphase systems, C p can be approximated by a weighted average of the component heat capacities:

Cp =



wi Cpi

(4.47)

334

TRANSPORT PROPERTIES OF MATERIALS

where wi is the weight fraction of each component. Since the relative heat capacities (per gram) of gas, polymer, and metal are 2:4:1, the heat capacity of polymer tends to dominate in these types of multicomponent systems, acting as a heat sink during processing. Due to the dependence on mean free path as described in Eq. (4.40), the thermal conductivity of heterogeneous systems is impossible to predict on heat capacity alone. As in previous sections, we do know that disorder tends to decrease thermal conductivity due to mean free path considerations, and this is indeed the case for fillers with high thermal conductivities, such as copper and aluminum in epoxy matrices (see Table 4.12). The thermal conductivity of the epoxy matrix increases only modestly due to the addition of even high percentages of thermally conductive fillers. One way to deal with heterogeneous materials is to assume that the composite has an effective thermal conductivity, keff , which can then be used in heat transport expressions as previously performed for homogeneous materials. There are several expressions for keff that have been developed, depending upon the type and amount of reinforcing material. For spherical particles of small volume fractions, φ, we have keff 3φ =1+  (4.48) k1 + 2k0 k0 −φ k1 − k0 where k1 and k0 are the thermal conductivities of the sphere and matrix phase materials, respectively. For large volume fractions, the thermal conductivity of the composite can be estimated with: keff 3φ  =1+  + 2k k1 − k 0 k k0 1 0 − φ + 1.569 φ 10/3 + · · · k1 − k0 3k1 − 4k0

(4.49)

Table 4.12 Thermal Conductivity of Epoxy Resin Filled with Various Compounds Thermal Conductivity (W/m·K) Filler

Vol% Filler

Filler

Epoxy Resin

Composite

Aluminum, 30 mesh Sand, coarse grain Mica, 325 mesh Alumina, tabular Alumina, 325 mesh Silica, 325 mesh Copper powder Alumina (tabular/325 mesh)

63 64 24 53 53 39 60 45/20

208 1.17 0.67 30.3 30.3 1.17 385 30.3

0.20 0.20 0.20 0.20 0.23 0.23 0.23 0.20

2.5 1.0 0.5 1.0 1.4 0.77 1.6 2.5

Source: F. Rodriguez, Principles of Polymer Systems, 2nd ed. Copyright  1982 by McGraw-Hill Book Co.

HEAT TRANSPORT PROPERTIES OF MATERIALS

335

Generally, the thermal interaction of spheres that can occur at high volume fractions is negligible, such that the simpler Eq. (4.48) is often used. For nonspherical reinforcement, such as aligned, continuous fibers, the thermal conductivity is anisotropic—that is, not the same in all directions. Rayleigh showed that the effective thermal conductivity along the direction of the fiber axis is keff ,z =1+ k0



k 1 − k0 k0



φ

(4.50)

whereas the effective thermal conductivity in the perpendicular directions are identical: keff ,y keff ,x 2φ  = =1+  k 1 − k0 k1 + k0 k0 k0 −φ+ (0.30584φ 4 + 0.013363φ 8 . . .) k1 − k0 k 1 − k0 (4.51) Additional expressions exist for nonspherical reinforcements and for solids containing gas pockets. The interested reader is referred to reference 15 for more information. 4.2.4.2 Thermal Conductivity of Laminar Composites. In the case of laminar composites or layered materials (cf. Figure 1.74), the thermal conductance can be modeled as heat flow through plane walls in a series, as shown in Figure 4.36. At steady state, the heat flux through each wall in the x direction must be the same, qx , resulting in a different temperature gradient across each wall. Equation (4.2) then becomes dT dT dT qx = −kA = −kB = −kC (4.52) dxA dxB dxC

where the minus sign simply indicates the direction of heat flow, depending upon how the temperature gradient is defined. For the case of constant thermal conductivity with temperature, which is true for small temperature gradients, and a constant cross-sectional heat flow area, A, the heat flow in the x direction, Qx = qx · A becomes (T2 − T3 ) (T3 − T4 ) (T1 − T2 ) = kB A = kC A (4.53) Qx = kA A xA xB xC

T1

A

B

C

T2

q

T3 T4 ∆xA

Figure 4.36

∆xB

∆xC

Heat flow through a multilayered wall.

336

TRANSPORT PROPERTIES OF MATERIALS

Example Problem 4.1

Question: Derive an expression analogous to Eq. (4.54) for steady-state heat flow through concentric cylinders, as shown at right. All cylinders are of the same length, L. Assume that thermal conductivity is constant with temperature, but that the material in each concentric ring is different. Answer : Fourier’s Law in cylindrical coordinates can be written as r2

dT Q = −k q= A dr

r3 r4

r1 q

The cross-sectional area normal to heat flow, A, is not constant, and varies with radius as A = 2πrL, where L is the length of the cylinders. For the inner cylinder, substitution of the area expression into Fourier’s Law and separation of variables gives QA 2πL



r2

r1

dr = −kA r



A B

T1

T2

T3

T4

C

T2

dT

T1

and integration results in QA =

kA 2πL (T1 − T2 ) ln(r1 /r2 )

Multiply the numerator and denominator by (r2 − r1 ), and defining a log-mean area, Alm as follows AA,lm =

(2πLr2 ) − (2πLr1 ) A2 − A1 = ln(2πLr2 /2πLr1 ) ln(A2 /A1 )

the heat flow through the inner cylinder becomes Q=

(T1 − T2 ) (T1 − T2 ) kAA,lm (T1 − T2 ) = = (r2 − r1 ) (r2 − r1 /kAA,lm ) RA

where RA is the resistance, which for the cylindrical geometry is defined as RA =

(r2 − r1 ) kA AA,lm

Finally, we recognize that at steady state, the heat flow through all the cylinders must be the same, Q = QA = QB = QC , so that Q=

(T1 − T2 ) (T2 − T3 ) (T3 − T4 ) = = RA RB RC

HEAT TRANSPORT PROPERTIES OF MATERIALS

337

As was the case for the multilayer wall, the internal wall temperatures T2 and T3 can be eliminated through combination of the three equations to yield one equation Q=

(T1 − T4 ) (T1 − T4 ) =  RA + RB + RC Ri

Note that this is the same result as Eq. (4.54), except that the resistance is defined differently.

Each of the three relations in Eq. 4.53 can be solved for the temperature difference, and the equations can be added to eliminate the internal surface temperatures, T2 and T3 , such that only the external surface temperatures, T1 and T4 remain. The heat flow is then expressed as follows: Qx =

(T1 − T4 ) (T1 − T4 ) = xA xB xC RA + RB + RC + + kA A kB A kC A

(4.54)

Cooperative Learning Exercise 4.9 Work in groups of three. Your plant is replacing asbestos used to insulate low-pressure steam pipes with a new multilayer composite material composed of a 2.5-cm-thick, polystyrene foam (k = 0.029 W/m·K) and a 2-mm-thick, outer, protective layer of polyethylene (k = 0.33 W/m·K). Assume that the pipe carries saturated steam at 130◦ C and that the outer surface of the insulation is 30◦ C. The piping is stainless steel (k = 14 W/m·K), with ID = 0.0254 m and OD = 0.0508 m. Use the equations developed in Example Problem 4.1 to solve the following problems, assuming a pipe length of 1 m. Person 1 : Calculate the log-mean heat transfer area for the stainless steel pipe, then use this to calculate the thermal resistance through the pipe. Person 2: Calculate the log-mean heat transfer area for the polystyrene foam only, then use this to calculate the thermal resistance through the foam. Person 3 : Calculate the log-mean heat transfer area for the outer polyethylene layer, then use this to calculate the thermal resistance through the polyethylene. Combine your resistance to calculate the heat transfer rate through the 1-m section of pipe. Which of the resistances dominates the calculation?

Answers: Ass,lm = 0.115 m2 ; Rss = 0.008 K/W APS,lm = 0.229 m2 ; RPS = 3.76 K/W APE,lm = 0.323 m2 ; RP E = 0.019 K/W Q = 26.4 W; polystyrene foam dominates

where Ri = xi /(ki ·A) is called the thermal resistance, as introduced at the end of Section 4.2.2.1.  The total thermal resistance is simply the sum of the individual resistances, Rtot = Ri . These “R values” are often reported in lieu of thermal con-

338

TRANSPORT PROPERTIES OF MATERIALS

ductivity values for common insulating materials, such as fiberglass and rock wool, for which thickness and sizes have been standardized, typically 1-ft2 area and 1-inch thickness. Selected thermal conductivities for these and other insulating materials can be found in Appendix 5.

4.2.5

Heat Transport Properties of Biologics

Heat storage in biological systems is much the same as for the previous material classes. For example, the heat capacity of biological fluids, which are heterogeneous suspensions and mixtures, can be approximated by Eq. (4.47). In instances where the biological fluid or food substance is an aqueous-based suspension or solution of fat and solids, the specific heat (heat capacity per unit weight in J/g·K) can be approximated as Cp = 4.2(xH2 O + 0.5xfat + 0.3xsolids )

(4.55)

where xi is the mass fraction of water, fat, and solids, respectively. Heat transfer in biological systems, on the other hand, is often a complex combination of conduction, convection, and radiation. For example, the insulating value of fur and feathers in animals is certainly measurable, but is based on more than just the thermal conductivity of the biological materials constituting the fur or feathers, since gases become entrapped in these often times porous materials, and heat is being radiated to and from the surroundings. As a result, the description of heat transfer in biological systems is much more empirical, and it involves a separate set of terminologies from the traditional, engineering-oriented set of terms. Nonetheless, the principles are the same, and we relax our restriction on heat transfer via conduction only so as to better describe heat transfer in “biological units”—animals and plants. 4.2.5.1 Thermal Conductance in Biological Systems. As introduced at the end of Section 4.2.2.1, thermal conductance is the thermal conductivity divided by the thickness (C = k/L), and thermal resistance is the inverse of the product of thermal conductance and area (R = 1/C·A). These terms are widely used to describe the insulating value of skin coverings such as fur and clothing in biological systems. As mentioned in the introduction to this section, the insulating value of fur and clothing is derived from more than thermal conductivity considerations. It involves the reduction of permeability to water vapor transmission, the disruption of convective heat transport from the surface, and an increase in surface area. Nonetheless, the insulating value of clothing has been standardized in terms of thermal conductance, and is expressed in units of clo. A clo is the amount of insulation that would allow 6.45 N·m/s of heat to transfer to the surroundings from a 1.0 m2 area of skin across a 1◦ C temperature difference between the skin and the environment. The higher the clo value of clothing, the less heat will be transferred from the skin. We will continue to use common engineering units for our analysis. For clothing, Fourier’s Law [Equation (4.2)], can be expressed as

qy =

dT Q = −kcl A dy

(4.2)

HEAT TRANSPORT PROPERTIES OF MATERIALS

339

which at steady state and for constant thermal properties, can be expressed in terms of conductance of clothing, Ccl , as Q=

−kcl A T = Ccl A(Tsk − T∞ ) y

(4.56)

where Tsk is the temperature of the skin and T∞ is the temperature of the surrounding environment. Some representative values for conductance of common items of clothing are given in Table 4.13. Clothing is often worn in layers, and the principles represented by Eq. (4.54)—that is, the additivity of resistances—apply here as well. Due to the added complexities of layered clothing in terms of simultaneous series and parallel heat conduction, the following empirical relationships have been developed for the thermal resistance of clothing, Rcl (in K/W), in men and women:  0.494 (1/Ccl ) + 0.0119 Rcl,men = (4.57) A  0.524 (1/Ccl ) + 0.0053 (4.58) Rcl,women = A where A is the surface area in m2 . Clothing surface area is based on a 15–25% increase over the surface area of a nude body, A (in m2 ), as determined by the formula of DuBois: (4.59) A = 0.07673W 0.425 Ht0.725 = 0.2025m0.425 Ht0.725 where W is the body weight in newtons, Ht is the height in meters, and m is the body mass in kilograms. Mean skin temperature, Tsk , is also affected by the presence of clothing. The following empirical relationship can be employed and is essentially independent of metabolic rate up to four or five times the rate at rest. For clothed humans, Tsk = 25.8 + 0.267T∞ (4.60) where T∞ is again the surrounding environmental temperature, in degrees Celsius. Table 4.13 Conductance Values for Selected Pieces of Clothing Clothing

Ccl (W/m2 ·K)

Undergarments Thermal undergarments Shirts, vests, and blouses Pants and Slacks Sweaters and jackets Sock Stockings, panty hose Sandals Boots

34–130 18 22–46 15–25 17–38 65–161 640 320 81

Source: A. T. Johnson, Biological Process Engineering. Copyright  1999 by John Wiley & Sons, Inc.

340

TRANSPORT PROPERTIES OF MATERIALS

Fur on animals is treated much the same way as clothing on humans. Some typical values of conductance for furs and feathers are presented in Table 4.14. The total surface area of an animal, A (in m2 ), can be estimated by A = 0.09m2/3

(4.61)

where m is the animal’s body mass, in kilograms. The values in Table 4.14 include the effects of convection and conduction, although radiation was not a significant contributor in these instances. We will see in the next section how the effects of convection can be combined with thermal conductivity to describe heat transfer in these systems. 4.2.5.2 Convective Heat Transfer in Biological Systems∗ . Unlike the materials of the previous sections, for which processing temperatures, thermal conductivities, and processing environments vary greatly, even within a material subclass (e.g., the heat transport processes involved in creating SiC via chemical vapor deposition are much different from those involved in forming SiC by a solid-state reaction), biological materials generally function within a relatively narrow temperature range, have thermal conductivities within an order of magnitude of each other, and function primarily in only one gaseous environment—air. As a result, we are able to briefly discuss here the role of convection in heat transfer in biological systems. To be sure, convection plays an important role in the processing and application of all the other material classes as well, and as you probably are already aware, the dimensionless groups and principles discussed here are general to heat transport in all systems. We will revisit some of these topics in Chapter 7, when the processing of materials is described.

Table 4.14 Selected Values of Thermal Conductance for Fur and Feathersa

Animal

Condition

C m Thickness (kg) (mm) (W/m2 ·K)

Rabbit Back 2.5 Horse Flank 650 Pig Flank 100 Chicken Well-feathered 2 Cow — 600 Dog Winter 50 Polar bear In air 400 Polar bear In water 400 Beaver In air 26 Beaver In water 26 Squirrel Summer — Monkey Summer — a

9 8 5 25 10 44 63 57 44 40 3 8

2.4 4.1 2.8 1.6 3.2 1.2 1.2 27 1.2 14 9.0 3.2

Includes skin and still air layer. Source: A. T. Johnson, Biological Process Engineering. Copyright  1999 by John Wiley & Sons, Inc.

HEAT TRANSPORT PROPERTIES OF MATERIALS

341

Convection involves the transfer of heat by means of a fluid, including gases and liquids. Typically, convection describes heat transfer from a solid surface to an adjacent fluid, but it can also describe the bulk movement of fluid and the associate transport of heat energy, as in the case of a hot, rising gas. Recall that there are two general types of convection: forced convection and natural (free) convection. In the former, fluid is forced past an object by mechanical means, such as a pump or a fan, whereas the latter describes the free motion of fluid elements due primarily to density differences. It is common for both types of convection to occur simultaneously in what is termed mixed convection. In such instance, a modified form of Fourier’s Law is applied, called Newton’s Law of Cooling, where the thermal conductivity is replaced with what is called the heat transfer coefficient, hc : qy =

Qy = hc dT A

(4.62)

Unlike Fourier’s Law, Eq. (4.62) is purely empirical—it is simply the definition for the heat transfer coefficient. Note that the units of hc (W/m2 ·K) are different from those for thermal conductivity. Under steady-state conditions and assuming that the heat transfer area is constant and hc is not a function of temperature, the following form of Eq. (4.62) is often employed: Qy = hc A T

(4.63)

where T is generally taken as the difference between the solid surface temperature and the fluid temperature far from the surface (bulk fluid temperature). Heat transfer coefficients cannot, in general, be calculated from first principles, and are normally derived from experiments with specific materials, fluids, geometries, and conditions. The use of dimensionless groups helps simplify this analysis. A complete review of dimensionless groups and their use in heat transfer analysis is not appropriate here. We will limit our discussion to three appropriate dimensionless groups: the Reynolds number, Re; Prandtl number, Pr; and Nusselt number, Nu. The Reynolds number is a ratio of inertial forces to viscous forces and is given by Re = Dvρ/µ, where D is the characteristic diameter in which a fluid is flowing, v is the fluid velocity, µ is the fluid viscosity, and ρ is its density. The Prandtl number was introduced in Section 4.2.1.2 as a ratio of momentum diffusivity to thermal diffusivity and is defined as Pr = Cˆ p µ/k, where µ is again the fluid viscosity, k is its thermal conductivity, and Cˆ p is the heat capacity. The Reynolds and Prandtl numbers provide a means for correlating measurable fluid properties (µ, ρ, k, and Cˆ p ) and geometric considerations (D, v) with the heat transfer coefficient, hc , in terms of the Nusselt number, Nu, which is defined as hc D (4.64) Nu = k where D is again a characteristic diameter and k is the thermal conductivity. Most correlations involve Pr and Re raised to an appropriate power: Nu = constant Ren Prm

(4.65)

342

TRANSPORT PROPERTIES OF MATERIALS

Example Problem 4.2 (Adapted from Johnson)

The metabolic rate of animals, M (in W), can be approximated by M = 3.39m0.75 , where m is the body mass of the animal in kilograms. About 15% of this heat is lost from the respiratory system, and another 50% is lost from the legs, tail, and ears. The remainder must be transported away from the body through the skin. Estimate the thermal conductivity of a 5-mm-thick hair coat of a 600-kg cow if the skin surface temperature is at 32◦ C and the ambient temperature is 15◦ C. Assume that the cow is standing still in a field, with a breeze of 0.5 m/s. For flow across cylinders where 1000 < Re < 50,000, the following correlation for Nusselt number may be used: Nu = 0.26(8173)0.6 (0.7)0.3 Answer : The heat generated by the cow is M = 3.39(600 kg)0.75 , so Q = 0.35(411) = 144 N · m/s The total surface area of the cow, as given by Eq. (4.61) is A = 0.09(600 kg)2/3 = 6.40 m2 We might estimate that about 20% of this surface area is in the legs and tail, which have already been accounted for in terms of heat loss, so that the remainder, about 5.12 m2 , is the surface area of the body. We can further model the cow’s body as roughly cylindrical in shape and can estimate that its radius is 40 cm. This allows us to calculate the length of the cylinder (including the surfaces at both ends) as A = 2πrL + 2πr 2 = 2π(0.4 m)L + 2π(0.4 m)2 = 5.12 m2 ⇒ L = 1.64 m We will assume forced convection in order to calculate dimensionless groups. Physical properties of air must be estimated, and we will use the ambient air temperature to estimate them. The following physical properties can be found in tables or calculated: ◦

ρ(15 C) = 1.226 kg/m3 , Cp = 1.0048 kJ/kg·K,



µ(15 C) = 1.8 × 10−5 kg/m·s, k = 0.0253 W/m·K

Pr = (1004.8 J/kg·K)(1.8 × 10−5 kg/m·s)/(0.0253 J/m·K·s) = 0.7 Re = (0.8 m)(0.5 m/s)(1.226 kg/m3 )/(1.8 × 10−5 kg/m·s) = 27244 Nu = 0.26(27, 244)0.6 (0.7)0.3 = 107 = hc D/k Solving for hc from definition of Nu (see Eq. 4.64), we obtain hc = 107(0.0253 J/m·K·s)/(0.80 m) = 2.17 W/K The total thermal resistance comes from a combination of forced convection from the body and thermal conduction through the skin: Rtot = Rconduction + Rconvection = Lhair /khair A + 1/ hc A = T /Q

MASS TRANSPORT PROPERTIES OF MATERIALS

343

Substituting values from above and solving for khair gives (0.005 m)/khair (5.12 m2 ) + 1/(2.17 W/K)(5.12 m2 ) = (32 − 15 K)/(144 N·m/s), khair = 0.035 W/m·K

The values of m and n vary greatly depending on whether the flow is turbulent or laminar and depending on the specific geometry involved in heat transfer—for example, flow from smooth-walled tubes, or non-Newtonian fluid flow. We will not review all of these correlations here. We simply note that with proper descriptions of the surrounding fluid and the geometry of the biological system, values for heat transfer coefficients can be estimated and used in the analysis of heat transfer in biological materials. In particular, these correlations are useful in estimating physical properties such as thermal conductivity in biological systems where direct measurement may be difficult, as illustrated in Example Problem 4.2.

4.3

MASS TRANSPORT PROPERTIES OF MATERIALS∗

We bring the analogies of momentum, heat, and mass transfer in materials to a close with a description of mass transport, often called mass diffusion, or simply diffusion. Many of the phenomena we have already studied depend upon diffusion. For example, most phase transformations (excluding “diffusionless” transformations such as martensitic) rely on the movement of atoms relative to one another. Corrosion cannot occur unless there is a movement of atoms or ions to a surface that allows the appropriate chemical reactions to take place. And, as we will see in Chapter 7, some processing methods such as chemical vapor deposition are wholly dependent upon the transport of chemical reactants to the location where reaction conditions allow products to form. This is why the term “diffusion-controlled” is often used to describe the kinetics of a process—the reaction is rapid relative to the time it takes to transport reactants to and/or products from the reaction site. For these reasons, it is important to study diffusion in materials, even though the driving forces are complex and the mass transport properties of a material are highly dependent upon the diffusing species of interest. Unlike momentum and heat transport, the entity being transported is not the same in all diffusion situations. We must describe not only the material in which the transport is taking place, but also what is being transported (see Appendix 6, for example). The movement of hydrogen gas through metals, for example, is much different from the movement of liquid water through metals. As a result, the systems we will address will be highly specific and not at all general. Nonetheless, we can begin as in the previous sections of this chapter, with a description of mass transport properties from a fundamental viewpoint, and extend this to liquids and solids, with the recognition that

344

TRANSPORT PROPERTIES OF MATERIALS

as systems become more complex, theoretical predictions must give way to empiricisms and correlations. 4.3.1

Mass Transport Properties of Metals and Alloys

Recall from the beginning of this chapter that the transport of mass at steady state is governed by Fick’s Law, given in one-dimension by Equation (4.4): ∗ = −DAB JAy

dcA dy

(4.4)

∗ is the molar flux of component A in the y direction, dcA /dy is the concentrawhere JAy tion gradient of A in the y direction, and DAB is the molecular diffusivity, or diffusion coefficient, of A in component B. As in the previous sections, we will concentrate on the material-related properties of transport, which in this section is the molecular diffusivity. We begin with a description of the molecular origins of mass diffusivity, and we then proceed to correlations and predictive relationships for the specific material classes.

4.3.1.1 The Molecular Origins of Mass Diffusivity. In a manner directly analogous to the derivations of Eq. (4.6) for viscosity and Eq. (4.34) for thermal conductivity, the diffusion coefficient, or mass diffusivity, D, in units of m2 /s, can be derived from the kinetic theory of gases for rigid-sphere molecules. By means of summary, we present all three expressions for transport coefficients here to further illustrate their similarities. √ −5 MT µ = 2.6693 × 10 (4.6) σ 2 √ T /M (4.34) k = 8.32 × 10−2 2 σ   3 −7 T /M D = 2.6280 × 10 (4.65) P σ 2

As before, M is the molecular weight of the rigid sphere, T is the absolute temperature, and σ and  are the same Lennard-Jones parameters used in Eq. (4.6) (note that technically,  = 1 for the rigid-sphere model). The pressure, P , in Eq. (4.65) is in units of atmospheres. Equation (4.65) gives the so-called self-diffusion coefficient, or diffusivity for a sphere diffusing through a gas of the same spheres (i.e., DAA ). Fick’s Law [cf. Eq. (4.4)] related the molar flux to the concentration gradient in a mixture of gases, A and B. In this case, the diffusion coefficient for the rigid-sphere model is −7

DAB = 2.6280 × 10

 T 3 (MA + MB )/2MA MB 2 P σAB AB

(4.66)

where MA and MB are the molecular weights of species A and B, respectively; σAB = 1 (σ + σB ), and AB is again approximately unity for rigid spheres. For rigid sphere 2 A gases, we note that the diffusivity of species B through A should be the same as the diffusivity for species A through B, that is, DAB = DBA .

MASS TRANSPORT PROPERTIES OF MATERIALS

345

4.3.1.2 Mass Diffusivity in Liquid Metals and Alloys. The hard-sphere model of gases works relatively well for self-diffusion in monatomic liquid metals. Several models based on hard-sphere theory exist for predicting the self-diffusivity in liquid metals. One such model utilizes the hard-sphere packing fraction, PF, to determine D (in cm2 /s):  πRT 1/2 (1 − P F )3 D = σ CAW (4.67) M 8P F (2 − P F )

where T and M are the absolute temperature and molecular weight, as before; R is the gas constant, σ is the collision diameter of the previous equations, and CAW , the Adler–Wainwright correction factor, is a function of PF that can be determined from Figure 4.37. The packing fraction can be calculated using Eq. (4.8) as before. Recall from Table 4.1 that PF = 0.45 near the melting point. The values predicted by Eq. (4.67) for several liquid metals are compared with experimental values in Table 4.15. Several empirical expressions also exist, including one that correlates self-diffusivity (in cm2 /s) with melting point, Tm : −6

D = 3.5 × 10



Tm M

1/2

Vm1/3

(4.68)

where M is the molecular weight and Vm is the atomic volume at the melting point. The self-diffusivity values predicted by Eq. (4.68) for several liquid metals at their melting point are compared in Table 4.15 with values from the semiempirical hard sphere model of Eq. (4.67) and experimental values. Notice the similarity in form of Eq. (4.65) with Eq. (4.7), which used the packing fraction to estimate the viscosity. In this case, however, the dependences are the inverse of packing fraction and molecular weight to those in Eq. (4.7). In other words, whatever factors that change viscosity will have the opposite effect on diffusion. As viscosity goes up, diffusivity goes down; the inverse is true as well. This relationship is stated

1.5 1.3

CAW

1.1 0.9 0.7 0.5 0.3

0

0.1

0.2

0.3

0.4

0.5

Packing fraction

Figure 4.37 The Adler–Wainwright correction, CAW , as a function of packing fraction, PF. Reprinted, by permission, from T. Iida and R. I. L. Guthrie, The Physical Properties of Liquid Metals, p. 209. Copyright  1988 by Oxford University Press.

346

TRANSPORT PROPERTIES OF MATERIALS

Table 4.15 Comparison of Experimental Melting Point Self-Diffusivities (in Units of 10−9 m2 /s) with Correlations for Selected Liquid Metals Metal Li Na K Cu Zn Ga Rb Ag Cd In Sn Hg Pb Al Ni Mg Sb Cs Fe

Experimental

Eq. (4.67)

Eq. (4.68)

5.76–6.80 3.85–4.23 3.59–3.76 3.97 2.03–2.06 1.60–1.71 2.22–2.68 2.55–2.56 1.78 1.68–1.69 2.05–2.31 0.93–1.07 1.68–2.19 — — — — — —

7.01 4.24 3.85 3.40 2.55 1.73 2.68 2.77 2.00 1.77 1.96 1.07 1.67 — — — — — —

6.72 4.10 3.71 3.18 2.45 1.64 2.58 2.68 1.94 1.72 1.86 0.93 1.60 4.87 3.90 5.63 2.66 2.31 4.16

Source: T. Iida, and R. I. L. Guthrie, The Physical Properties of Liquid Metals. Copyright  1988 by Oxford University Press.

explicitly in the Stokes–Einstein equation D=

kB T 4πµr

(4.69)

where kB is Boltzmann’s constant, T is absolute temperature, µ is the viscosity of the “solvent” (in this case, the molten metal), and r is the radius of the diffusing hard sphere (in this case, a molten metal atom). Equation (4.69) is the Stokes–Einstein equation for self-diffusion. Another form of the Stokes–Einstein equation can be developed for dilute solutions of large spheres in small, spherical molecules, which leads to a binary diffusion coefficient: kB T DAB = (4.70) 6πµr Equation (4.70) is a starting point in the determination of diffusivities in liquid metal alloys, but in most real systems, experimental values are difficult to obtain to confirm theoretical expressions, and pair potentials and molecular interactions that exist in liquid alloys are not sufficiently quantified. Even semiempirical approaches do not fare well when applied to liquid alloy systems. There have been some attempts to correlate diffusivities with thermodynamic quantities such as partial molar enthalpy and free energy of solution, but their application has been limited to only a few systems.

347

MASS TRANSPORT PROPERTIES OF MATERIALS

4.3.1.3 Mass Diffusivity in Solids. Fick’s Law [Eq. (4.4)] implies that a concentration gradient, dc/dy, is necessary for diffusion and molar flow to occur. In solid metals such as pure nickel, no concentration gradient exists, yet radioactive tracer experiments tell us that diffusion still takes place. As illustrated in Figure 4.38, a thin layer of radioactive nickel placed at one end of a pure nickel bar will slowly diffuse through the pure metal, a process that can be monitored easily by radiation detection. This tells us that even though no concentration gradient exists (the chemical potentials of radioisotopes are identical, hence there is no difference in concentration across the interface), there is still atomic movement. We know that above absolute zero, atomic vibrations occur in solids. It is these small amplitude vibrations that allow some atoms to overcome an activation energy barrier and move from one position to another in the lattice. This type of random walk process allows atomic interchange to take place—neighboring atoms simply switch places due to thermal vibrations. Experiments such as the one illustrated in Figure 4.38 not only give us self-diffusion coefficients for certain substances, but as the temperature of the experiment is varied, they give us the temperature dependence of the process and a measurement of the activation energy barrier to diffusion. Diffusion in solid systems, then, can be modeled as an activated process; that is, an Arrhenius-type relationship can be written in which an activation energy, Ea , and temperature dependence are incorporated, along with a preexponential factor, D0 , sometimes called the frequency factor:

D = D0 exp(−Ea /RT )

(4.71)

Diffusion zone

Nonradioactive nickel

d (a)

Amount of radioactivity

Nonradioactive nickel

Radioactive nickel

0

0

x (b)

Figure 4.38 Schematic illustration of a self-diffusion experiment in which (a) a thin layer of radioactive nickel is deposited on one surface of a nonradioactive nickel specimen. After heating and time (b), the radioactive nickel has diffused into the sample, as monitored with a radioactive detector. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

348

TRANSPORT PROPERTIES OF MATERIALS

Since this is an empirical expression, it can be applied to almost any diffusion process, but it does not tell us much about the diffusion mechanism. In general, three types of processes have been identified for diffusion in solids: an interchange mechanism (sometimes called an exchange mechanism), a vacancy mechanism, and an interstitial mechanism. These mechanisms are illustrated in Figure 4.39. A fourth mechanism, ring interchange, is possible, but rarely observed, and will not be considered here. In most metals, the vacancy mechanism (Figure 4.39a) dominates, because it requires a relatively small amount of energy, and there are a sufficient number of vacancies in most metals. The interstitial mechanism (Figure 4.39b) is possible in metals since a self-interstitial is a relatively mobile species requiring shorter distances of atomic movement, but the equilibrium number of self-interstitials is relatively small at almost all temperatures. Finally, the direct interchange mechanism (Figure 4.39c) is possible in metals, but requires an extremely high activation energy. Activation energies for self-diffusion are expected to be in the range of 232–261 kJ/mol for the vacancy mechanism and 492–615 kJ/mol for the interstitial mechanism. Observed values are typically in the range of 164–203 kJ/mol. Some representative values for self-diffusion activation energies, as well as frequency factor in Eq. (4.71), D0 , are given in Table 4.16. Bear in mind that D0 contains the temperature-independent portions of diffusion, such as entropic effects, whereas Ea reflects all activation energy processes, including both atomic mobility and vacancy formation in the case of diffusion via the vacancy mechanism. Notice in Table 4.16 that there is a general trend of increasing activation energy for diffusion with increasing

(a)

(b)

(c)

Figure 4.39 Illustration of (a) vacancy, (b) interstitial, and (c) interchange (exchange) mechanisms in atomic diffusion. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

MASS TRANSPORT PROPERTIES OF MATERIALS

349

Table 4.16 Frequency Factors, Activation Energies and Melting Points for Self-Diffusion in Selected Metals Species Pb Zn (|| to c axis) Zn (⊥ to c axis) Ag Au Cu Ni Co α-Fe γ -Fe Pt

D0 (cm2 /s)

Ea (KJ/mol)

Tm (◦ C)

6.6 0.046 91 0.89 0.16 11 — 0.31 2300 5.8 0.048

117 85 130 192 222 239 293 280 306 310 233

327 419 419 962 1064 1085 1455 1494 1538 — 1772

Source: B. Chalmers, Physical Metallurgy. Copyright  1959 by John Wiley & Sons, Inc.

melting point. Recall from Chapter 1 that intermolecular forces have a profound influence on a number of physical properties, such as melting point. The more strongly atoms are attracted to one another, the more energy it requires to pull them apart. This is reflected in both the melting point and the activation energy for diffusion, in which atoms must overcome attractive forces in order to move away from one another. In alloys, attractive forces still play an important role, but the situation is slightly different insofar as solute atoms are sufficiently small relative to the host atom to allow the interstitial mechanism to dominate (see Figure 4.40). As a result, hydrogen, carbon, nitrogen, and oxygen diffuse interstitially in most metals with relatively low activation energy barriers. The diffusion of carbon in iron has been particularly well-studied in conjunction with the phase transformations in the C–Fe system (cf. Figure 2.8). An example of the excellent fit provided by Equation (4.71) to experimental data in the C–Fe system is shown in Figure 4.41. Notice how the activation energy for this interstitial diffusion process is about 85 kJ/mol, which is much smaller than for vacancy-driven diffusion, due to the small size of the carbon atoms relative to iron. Not only do larger atoms, such as other metals, have difficulty diffusing interstitially, but also atomic interactions between dissimilar atoms greatly affect the activation energy for diffusion. In addition to concentration and random walk diffusion, which are termed ordinary diffusion, there are other types of diffusion that can be important under certain circumstances. Atomic mobility can be brought about by thermal and pressure gradients. There are also such structurally driven diffusion processes as dislocation diffusion, surface diffusion, and grain-boundary diffusion. The latter is particularly important in polycrystalline materials where grain boundaries predominate. The grain boundaries provide a path of lower activation energy for movement than through the bulk solid. Moreover, grain boundary diffusion is less temperature-variant than bulk, or volume, diffusion in these materials, as illustrated for single crystalline and polycrystalline silver in Figure 4.42.

350

TRANSPORT PROPERTIES OF MATERIALS

(a)

(c) (b)

(d)

Figure 4.40 Illustration of diffusion mechanisms in alloys and ionic solids: (a) interchange (exchange); (b) ring rotation (rare); (c) interstitial migration; and (d) vacancy migration. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

−40

−10

−27

100

200

300

400

500

800 700 600

T (°C)

−6 Log10 D =

−4400 T

−5.70

or

−10

log10D (m2/s)

D = 2 × 10−6 exp

−84,300 2 m /S RT

−14

−18

−22

1

2

3

4

1 × 103 T

Figure 4.41 Variation with temperature of the diffusivity for carbon in BCC-Fe (α − Fe). Activation energy is in units of J/mol. Reprinted, by permission, from D. R. Gaskell, An Introduction to Transport Phenomena in Materials Engineering, p. 519. Copyright  1992 by Macmillan Publishing Co.

MASS TRANSPORT PROPERTIES OF MATERIALS

351

Temperature (°C) 900

−12

800

700

600

500

450

400

Polycrystal

log D (m2/s)

Single crystal

−14

−16

−18 0.80

D = 2.3 × 10−9 exp −13,300 T

Dvol = 8.95 × 10−5 exp −23,100 T

1.00 103

1.20 1 −1 × (K ) T

1.40

1.60

Figure 4.42 Self-diffusion coefficient in single and polycrystalline silver, illustrating the effect of grain boundary diffusion, especially at lower temperatures. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

Cooperative Learning Exercise 4.10 Consider the diffusivity data presented in Figure 4.42 for the self-diffusion of silver in polycrystalline and single-crystalline form. Use the correlations provided to calculate the following quantities. Person 1 : Calculate the self-diffusivity in polycrystalline silver (grain boundary diffusion), DGB , at 500◦ C in m2 /s. What is the activation energy for this process in kJ/mol? Person 2 : Calculate the self-diffusivity in single-crystalline silver (volume diffusion), Dvol , at 500◦ C in m2 /s. What is the activation energy for this process in kJ/mol? Combine your answer to get the ratio of grain boundary diffusion to volume diffusion for silver at 500◦ C. How would this change at higher temperatures? How would this change as the grain size decreased further?

Answers: DGB (500◦ C) = 2.3 × 10−9 exp(−13, 300/773) = 7.8 × 10−17 m2 /s Ea = (13,300 K)(8.314 J/mol·K) = 111 kJ/mol Dvol (500◦ C) = 8.95 × 10−5 exp(−23, 100/773) = 9.4 × 10−18 m2 /s Ea = (23,100 K)(8.314 J/mol·K) = 192 kJ/mol DGB /Dvol = 7.8/0.94 = 8.3. We would expect this ratio to decrease at higher temperature, eventually approaching 1.0 as volume diffusivity increases due to its stronger temperature dependence. This ratio should get larger with decreasing grain size as there are more grain boundaries per unit volume.

352

TRANSPORT PROPERTIES OF MATERIALS

4.3.2

Mass Transport Properties of Ceramics and Glasses

As with thermal conductivity, we see in this section that disorder can greatly affect the mechanism of diffusion and the magnitude of diffusivities, so that crystalline ceramics and oxide glasses will be treated separately. Finally, we will briefly describe an important topic relevant to all material classes, but especially appropriate for ceramics such as catalyst supports—namely, diffusion in porous solids. 4.3.2.1 Diffusion in Crystalline Oxides. We wish to investigate the individual terms in Eq. (4.71) (i.e., D0 and Ea ), in more detail as they apply to crystalline oxide ceramics. To do so, we look first at a model ionic system, potassium chloride, KCl, realizing that the results will be general to all ceramics that have close-packed anion lattices and in which diffusion occurs by a vacancy mechanism. Diffusion of K+ ions in KCl occurs by interchange of the potassium ions with cation vacancies (see Figure 4.40d). It makes sense, then, that the diffusivity of potassium ions in KCl, DK,Cl is both a function of the potassium ion mobility, Di , and the cation vacancy concentration, [Vk′ ]: DK,Cl = Di [VK′ ] (4.72)

Let us first concentrate on the potassium ion mobility, Di , in Eq. (4.72). As with any activated process, the potassium ions must overcome an activation energy barrier, which we will notate as a free energy G† , to move to the cation vacancy site, as illustrated in Figure 4.43. Recall from Section 3.0.2 that the fraction of ions with the required energy to overcome the barrier is given by the Boltzmann distribution and depends on exp(−G† /kB T ). It can be shown, then, that Di = γ λ2 v exp(−G† /kB T )

(4.73)

where γ is a nearest-neighbor factor on the order of unity, λ is the jump distance, and v is the frequency factor, which for solids has a value of about 1013 s−1 . Turning now to the vacancy concentration in Eq. (4.72), recall from Chapter 1, and Eq. (1.48) in particular, that the concentration of vacancies is related to the free energy of formation of the vacancy, so that for a cation–anion vacancy pair, or Schottky defect, the concentration of vacancies, [Vk′ ] is given by nV /N = [Vk′ ] = exp(−Gs /2kB T )

(4.74)

where nV is the number of defects (vacancies), N is the total number of lattice sights, T is the absolute temperature, kB is Boltzmann’s constant, and Gs is the free energy change necessary to form a Schottky vacancy pair. We know that the free energy can be broken down into entropic and enthalpic components [Eq. (2.12)], such that   −Ss T + Hs ′ [Vk ] = exp(−Gs /2kB T ) = exp (4.75) 2kB T If we perform a similar breakdown of G† in Eq. (4.73) and substitute Eqs. (4.73) and (4.75) into (4.72), we obtain the following relationship for the diffusivity of potassium ions in KCl:     S † + Ss /2 −H † − Hs /2 exp DK ,Cl = γ λ2 v exp (4.76) kB kB T

MASS TRANSPORT PROPERTIES OF MATERIALS

353

Free energy

∆G †

∆m N λ

Distance

Figure 4.43 Diffusion in a potential gradient µ, where G† is the height of the activation energy barrier and λ is the jump distance. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

The important parts of Eq. (4.76) are the exponential terms. The first exponential, which contains the entropies associated with potassium ion movement and vacancy formation, respectively, form the temperature-independent contributions to D0 , as discussed in the previous section. The second exponential, which contains the enthalpies of the two processes and the temperature dependence, form the activation energy, Ea , and temperature dependence of Eq. (4.71). To complete our discussion on diffusion in KCl, then, Eq. (4.76) tells us that a plot of DK,Cl versus 1/T should result in a straight line with slope of ( H † /kB + Hs /2kB ). As shown in Figure 4.44, this is the case only at high temperatures, where the intrinsic properties of the material dominate. In the lower-temperature region, impurities within the crystal fix the vacancy concentration. This is called the extrinsic region, where the diffusion coefficient is a function of divalent cation impurity concentration, such as calcium. The insert in Figure 4.44 shows that the cation vacancy concentration becomes fixed at low temperatures, and this is reflected in the fact that the slope of the diffusivity curve becomes H † /kB only. The change in slope occurs at the point where the intrinsic defect concentration (attributable only to temperature) is comparable with the extrinsic defect concentration (attributable to impurities). When the Schottky formation enthalpy is in the range of 628 kJ/mol, as is typical for BeO, MgO, CaO, and Al2 O3 , the crystal must have an impurity concentration smaller than 10−5 before intrinsic diffusion can be observed at 2000◦ C. It is thus unlikely that intrinsic diffusion will be observed in these oxides because impurity levels of only parts per million are sufficient to control the vacancy concentrations.

354

TRANSPORT PROPERTIES OF MATERIALS

104 103

10−1

102

10−2

∆Hs 1.3eV = 2k k

10−3

[Vk] = [F k0]

Intrinsic Region log [Vk]

101 0

10−4

10−1

10−5

10−2

10−6

10−3

10−7 1.0

log Dk (cm2/sec)

2.0 1000/T

10−4 10−5 10−6 ∆H + ∆HS + kB 2kB

10−7 10−8 10−9

∆H + kB

10−10 10−11 10−12 10−13 10−14

Extrinsic Region

10−15 10−16 0.0

1.0

2.0

3.0

1000/T (°K)

Figure 4.44 Diffusivity as a function of temperature for KC1 with 10−5 atom fraction divalent cation impurities. Insert plot shows the variation of vacancies with temperature. Adapted from W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

The experimental diffusion coefficients for most oxide ceramics do not generally show this intrinsic/extrinsic behavior illustrated in KCl. This is probably due to the fact that diffusion data are normally collected over a relatively small temperature range, as illustrated for a number of oxides in Figure 4.45. However, some diffusion in oxides is clearly extrinsic, such as for the diffusivity of oxygen in oxides containing the fluorite structure, such as UO2 , ThO2 , and ZrO2 . In these compounds, the addition of divalent or trivalent cation oxides, such as CaO, has been shown to fix the concentration of oxygen

355

MASS TRANSPORT PROPERTIES OF MATERIALS

Temperature (°C) 1716

10−5

10−6

1393

1145

977

828

727

K in b − Al2O3

Co in CoO (air)

O in Ca0.14Zr0.86O1.86

Fe in Fe0.95O

Cu in Cu2O

10−7

(Po2 1.4 × 10−4 atm) O in Y2O3

Diffusion coefficient (cm2/s)

10−8

10−9

N in UN Mg in MgO (single) Y in Y2O3

Cr in Cr2O3

O in CoO Po2 = 0.2 atm

−10

10

Al in Al2O3

O in Cu2O (Po2 = 0.2 atm)

182.4

45.6 31.9 21.7 13.7 kcal/mole

O in Ni0.68Fe2.32O4 (single)

10−11 Ni in NiO (air)

O in Al2O3 (poly)

91.2

10−12 O in Al2O3 (single)

O in UO2.00 O in TiO2 (P o2 = 1 atm)

−13

10

C in graphite 10−14

O in Cr2O3 Ca in CaO U in UO2.00

10−15

0.4

0.5

0.6

O in fused SiO2 (Po2 = 1 atm)

O in MgO 0.7

0.8

0.9

1.0

1.1

1/T × 103 (K−1)

Figure 4.45 Diffusivities in some common ceramics as a function of temperature. The activation energy can be estimated from the slope of the insert. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

356

TRANSPORT PROPERTIES OF MATERIALS

vacancies and make their concentration independent of temperature. Consequently, the oxygen ion diffusivity has a temperature dependence related only to the activation energy associated with oxygen ion mobility (121 kJ/mol). Cation diffusion coefficients, on the other hand, such as Mg diffusion in MgO, are more difficult to categorize as wholly intrinsic or extrinsic. The relatively high activation energies in these systems suggest that high- temperature measurements may be intrinsic and independent of minor impurities, but it is difficult to generalize. Intrinsic diffusion behavior is dominant in nonstoichiometric oxides such as zinc oxide or cobaltous oxide in equilibrium with oxidizing or reducing atmospheres. In these systems, the interstitial and vacancy concentrations are a function of the partial pressure in the surrounding environment. The dependence on partial pressure is related to whether the nonstoichiometric oxide is metal-deficient or oxide-deficient. In metaldeficient oxides such as FeO, NiO, and MnO, the concentration of vacant cation sites can be very large. For example, Fe1−x O contains 5–15% vacant iron sites. In this case, the solubility of oxygen in the metal oxide becomes an issue, and the metal ion diffusivity, DM , can be a function of oxygen partial pressure, as illustrated in Figure 4.46a. The free energy of solution for oxygen, GO , must then be considered, which in turn leads to a functional dependence of DM and vacancy concentration on the enthalpy of solution for oxygen, HO , as shown in Figure 4.46b, whereas the

∆H † +

6

∆H † + 1 ∆H0 3 kB

log DM

log DM

kB

1

const T

∆ HS 2

const Po2 1/T

(a)

(b)

log [VM″]

∆H † 2kB

const Po2

∆H † 3kB

log BM

log Po2

∆H † kB

const Po2 1/T

1/T

(c)

(d)

Figure 4.46 Schematic representation of the variation of metal ion diffusivity as a function of (a) oxygen partial pressure and (b) temperature, and (c) the metal ion vacancy concentration and (d) mobility with temperature. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

∆H † + ∆HS 2kB ∆HO 3 kB

log DO

log [VO••]

MASS TRANSPORT PROPERTIES OF MATERIALS

kB

∆HS 2 ∆H † +

kB

∆HO 3

const Po2 1/T (a)

357

∆H † kB

1/T (b)

Figure 4.47 Schematic representation of temperature dependence of (a) oxygen vacancy and (b) oxygen diffusivity in oxygen-deficient oxide ceramics. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

metal ion mobility, BM , remains a function of the activation energy barrier only, as illustrated in Figure 4.46d. A similar situation arises in oxygen-deficient oxides, such as CdO or Nb2 O5 , except that now three temperature regimes are possible, as shown in Figure 4.47. In the low-temperature regime, oxygen vacancy concentration is controlled by impurities, and extrinsic behavior is observed. At intermediate temperatures, the oxygen vacancy concentration changes due to a change in oxygen solubility with temperature. This is nonstoichiometric behavior. At high temperatures, thermal vacancies become dominant, and intrinsic behavior dominates. Chemical effects also occur in crystalline oxides; that is, impurity atoms diffuse at varying rates through oxides. In all cases, cation diffusion is much faster than oxygen diffusion, but similar cations, such as Ca and Mg, can behave very differently in different oxides. This is due to differences in chemical potential and activity, and mobility differences. As in metals, dislocation, surface, and grain boundary diffusion can be important. Grain boundary diffusion, in particular, can lead to diffusion coefficients that are 103 to 106 times greater than the bulk diffusion coefficient. However, the effect is not the same for both cations and anions. In some systems, only the ion that is expected to be present in excess at the grain boundary will exhibit enhanced grain boundary diffusion. 4.3.2.2 Diffusion in Glasses—Solubility and Permeability. The principles developed in previous sections apply to noncrystalline solids as well, but the terminology used is often different, owing to the applications for which diffusion in oxide glasses are important. There are two such situations that will be described here: (a) diffusion of gases through glass and (b) diffusion of cations through glass. The former is important in many applications, such as high-vacuum use, where glass is a common material of construction. The latter is important in glass containers, where cations in the glass can leach out over time depending on the solution it contains. Yet another application, the diffusion of gases through molten glasses, is important from an industrial standpoint in the fining of glasses, but is beyond the scope of this text. In both gas and cation diffusion, the movement of atoms through solid glasses is often described in terms of the permeability, PM , which is directly proportional to the

358

TRANSPORT PROPERTIES OF MATERIALS

diffusivity, D: PM = SD

(4.77)

where S is the solubility. In this case, the permeability is defined as the volume of gas at standard temperature and pressure (STP) passing per second through a unit area of glass of thickness 1 cm, with a 1 atm pressure difference across the glass. The solubility is defined as the volume of gas at STP dissolved in a unit volume of glass per atmosphere of external gas pressure. The diffusivity is defined as before. It should be clear from Eq. (4.77) that determination of two of the parameters fixes the third. Solubility increases with temperature according to an Arrhenius-type relationship, where the heat of solution serves as the activation energy. Since diffusivity also follows an Arrhenius expression, as described in the previous sections, one would expect that the permeability would also show a similar temperature dependence, and this is indeed the case. An example of the logarithmic permeability dependence on inverse temperature is shown in Figure 4.48. The differences in permeabilities among the various glasses can be explained by considering the network-modifying cations as “hole blockers” in the glass network. Thus, the permeability would be expected to increase with an increase in the concentration of network-forming constituents such as SiO2 , B2 O3 , and P2 O5 (cf. Section 1.2.4). Such an increase has been shown experimentally. As for the diffusivity of gases in glasses, the activation energy increases with increasing size of the gas molecule, as illustrated in Figure 4.49. For these gases, the frequency factor, D0 , is in the range of 10−4 to 10−3 cm2 /s. 800

400

−8

200

+25

100

−25

0

−78

°C

Fused silica

log10 permeation velocity

−9 −10 −11 −12

Combustion tubing no. 1720

Vycor

X-ray shield glass

−13

Phosphate Lead borate

−14

Soda lime no. 008

−15 0.5

1.0

1.5

2.0

2.5

3.0

3.5

Boro-silicate Chemical no. 650 pyrex no. 7740

4.0

4.5

5.0

1000/T

Figure 4.48 Permeability of helium through various glasses. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

MASS TRANSPORT PROPERTIES OF MATERIALS

Activation energy (kcal/mol)

25

359

N

20

15 Ne 10

H

He 5

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 Radius of diffusing species(Å)

Figure 4.49 Activation energy for diffusion as a function of atomic radius of diffusing species in silica glass. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

In the case of oxygen transport, diffusion may take place either as molecular or network diffusion, although molecular diffusion apparently takes place some seven orders of magnitude faster than network diffusion. The movement of water through glasses can also be important, particularly in the dehydroxylation of hydrated glasses such as vitreous silica. The hydroxyl content has been shown to greatly affect the properties of glasses, including reductions in viscosity and glass transition temperatures, and increases in phase separation and crystallization rates with increasing hydroxyl content. Some data from a typical dehydroxylation experiment for vitreous silica are shown in Figure 4.50. Here, t is time and L is the sample thickness. The value of t 1/2 L at 50% hydroxyl removal is shown in the plot. This value is significant, because the diffusivity can be estimated according to the following relationship: D=

0.049 (t/L2 )50%

(4.78)

where (t/L2 )50% is the value of t/L2 at removal of exactly 50% of the hydroxyl groups from the glass. For the movement of cations in glasses, as the amount of modifying cations in silicate glasses is increased, the activation energy for their diffusive transport decreases, resulting in an increase in the diffusivity. The increase in cation concentration presumably leads to a break up of the network and a decrease in the average interionic separation. Divalent cations diffuse much more slowly at a given temperature than monovalent cations, and their activation energies are much larger. Diffusivities of modifying cations in rapidly cooled glasses are generally higher than those in well-annealed glasses. These differences, often as large as an order of magnitude, reflect differences in

360

TRANSPORT PROPERTIES OF MATERIALS

100 80 34,300 60 %

50% 40 20 0

0

25

50

75

100 × 103

√t/L (√s/cm)

Figure 4.50 Experimental dehydroxylation curve for vitreous silica at constant temperature. Reprinted, by permission, from J. E. Shelby, in Experimental Techniques of Glass Science, C. J. Simmons and O. H. El-Bayoumi ed. p. 374. Copyright  1993, The American Ceramic Society.

specific volumes of the glasses, with the larger-volume, more open structure possessing the larger diffusivity. The process of cation movement in glasses is important for a number of processes, such as ion exchange and dissolution. Many univalent cations in glasses can be substituted for one another, such as the complete substitution of Na ion by Ag ions, even at low temperatures. Ions such as Li+ and K+ can also be substituted for Na+ , although stresses can build up in the glass due to cation radius differences. This procedure can actually be utilized to strengthen glasses, as will be discussed in Chapter 5. In the dissolution of glasses, ions in solution attack the glass network through leaching of network modifiers, such as Na+ . This is particularly prevalent in acidic aqueous solutions, where the countermigration of H+ ions into the network accelerates the leaching process, although an exact 1:1 exchange of Na+ :H+ does not necessarily take place. Water is involved in the process as well. A theoretical description of this process, or even one in terms of diffusion theory, is difficult, since the composition of the diffusing matrix is changing with time. A simple formula for determining the amount of leachedout alkalis, [Q], is to assume a power-law dependence with time, [Q] = kt α , where α typically has values between 0.45 and 0.9. A square-root dependence on leaching time (i.e., α = 0.5) is predicted by theory, but is not necessarily experimentally observed. The properties and development of alkali- and hydroxyl-resistant glasses are important areas of investigation, and the interested reader is referred to the many fine texts on these subjects listed at the end of this chapter. 4.3.2.3 Diffusion in Porous Solids. To this point, we have primarily concentrated on the atomic-level movement of atoms involved in diffusion. There are macroscopic effects to consider as well. For example, liquids and gases moving through porous solids can often travel faster through the pores, which are also filled with either gas or liquid, than through the solid particles. If the pores are small relative to the mean free path of the diffusing species (Figure 4.51a), then the diffusion rate can be influenced strongly

361

MASS TRANSPORT PROPERTIES OF MATERIALS

l l l d PA1

PA 2

PA1

PA 2

(a)

Figure 4.51 Diffusion in (c) transitional diffusion.

PA1

PA 2

(b)

pores:

(c)

(a) Knudsen

diffusion,

(b) Fickian

diffusion,

and

by collisions with the pore walls in a process known as Knudsen diffusion. Knudsen diffusion is rare for diffusing liquids since their mean free paths are so small. It can be significant, however, for gases, especially at low pressures or high temperatures. Knudsen diffusion is characterized by the Knudsen number, Kn = l/d, where l is the mean free path of the diffusing species, and d is the pore diameter in the same units as l. For Kn < 0.01, Fickian diffusion occurs (Figure 4.51b). For Kn > 100, Knudsen diffusion dominates. In this case, a Knudsen mass diffusivity, DKA (in m2 /s), is defined as follows:  (4.79) DKA = 48.5d T /MA where d is again the pore diameter (in meters), T is the absolute temperature in degrees kelvin, and MA is the molecular weight of the diffusing species, A, in g/mol. It is interesting to note that for a multicomponent gas, DKA is completely independent of the diffusivity in the other gaseous species, since A collides with the walls of the pore and not with the other gases. Each gas will have its own Knudsen diffusion coefficient in this case. When Kn ∼ 1.0, neither Fickian nor Knudsen diffusion dominates (Figure 4.51c), and transition diffusion occurs. The diffusivity in this case can be estimated in a number of ways, including a vectorial combination of the two diffusion coefficients. In the specific instance of A diffusing through a catalytic pore and reacting at the end of the pore to form gaseous component B, equimolar counterdiffusion can be assumed, ′ , is independent of concentration and and an effective transition region diffusivity, DNA can be calculated from the Knudsen and binary diffusion coefficients: ′ DNA =

1/DAB

1 + 1/DKA

(4.80)

This simplified diffusivity is sometimes used for diffusion in porous catalysts even when equimolar counterdiffusion is not occurring. This greatly simplifies the equations. When no reactions are occurring, the diffusivity is a function of concentration (in terms of mole fraction, xA ): DNA =

1 (1 − αxA )/DAB + 1/DKA

(4.81)

362

TRANSPORT PROPERTIES OF MATERIALS

Cooperative Learning Exercise 4.11

A gas mixture at a total pressure of 0.10 atm absolute and 298 K is composed of N2 (component A) and He (component B). The mixture is diffusing through an open capillary having a diameter of 5 × 10−7 m. The mole fraction of N2 at one end is xA1 = 0.8 and at the other end is xA2 = 0.2. The mean free paths for nitrogen and helium at 298 K are 654 ˚ respectively (1 A ˚ = 10−10 m); and their collision diameters, σ , are 3.75 and and 1936 A, ˚ respectively. 2.18 A, Person 1 : Calculate the Knudsen number for N2 . Is the diffusion molecular, Knudsen, or transitional? Estimate the Knudsen diffusivity for nitrogen, DKA , under these conditions. Person 2 : Estimate the hard-sphere binary diffusivity for nitrogen in helium, DAB , under these conditions. You may assume that AB = 1.0 in Eq. (4.66). Combine your information to estimate the transitional diffusivity for nitrogen, DN′ A . You can use an average concentration for nitrogen in your calculation. How does it compare to the two diffusivities you calculated individually? How does it compare to the transitional diffusivity if there were equimolar counterdiffusion—that is reaction at the end of the pore? Answers: Kn = 0.13 for N DKA = 7.9 × 10−5 m2 /s DAB = 5.8 × 10−4 m2 /s 2 ′ −5 m2 /s; a difference of about 9% DN A = 6.3 × 10−5 m2 /s DN A = 6.95 × 10

where α is the flux ratio factor, which for flow in an open system is related to the molecular weights of the two components:

α =1−



MA MB

(4.82)

When chemical reaction occurs, α must be defined in terms of molar fluxes. For the case of equimolar counterdiffusion, α = 1, and Eq. (4.81) reduces to (4.80). In real porous solids, the pores are not straight, and the pore radius can vary. Two parameters are used to describe the diffusion path through real porous solids: the void fraction, ε, defined as the ratio of pore area to total cross-sectional area, and the tortuosity, τ , which corrects for the fact that pores are not straight. The resulting effective diffusivity is then DA,eff = DAB (ε/τ ) (4.83) Tortuosity values range from about 1.5 to over 10. A reasonable range of values for many commercial porous solids is about 2–6. 4.3.3

Mass Transport Properties of Polymers

4.3.3.1 Diffusion in Polymer Solutions. A dilute polymer in a low-molecular weight solvent can be modeled as bead-spring chain (see Figure 4.52a). Each chain is a linear arrangement of N beads and N − 1 Hookean springs (see Chapter 5). The beads are characterized by a friction coefficient, represented by the lowercase Greek letter zeta, ζ , which describes the resistance to bead motion through the solvent. The model also allows for hydrodynamic interactions between the beads and the solvent

MASS TRANSPORT PROPERTIES OF MATERIALS

363

(a)

(b)

Figure 4.52 Single-molecule bead spring models for (a) dilute polymer solution and (b) polymer melt. From R. B. Bird, W. E. Stewart, and E. N. Lightfoot, Transport Phenomena, 2nd ed. Copyright  2002 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

molecules. The general result, which holds true for most solutions, relates the diffusivity to the friction coefficient: kB T DAB = (4.84) ζ The form of Eq. (4.84) should look familiar. For large, spherical particles in a lowmolecular-weight solvent, ζ = 6πµr, where µ is the viscosity of the pure solvent and r is the large particle radius, and Eq. (4.84) becomes Equation (4.70) a form of the Stokes-Einstein equation, which gives the binary diffusion coefficient, DAB , DAB =

kB T 6πµr

(4.70)

Equation (4.69) for self-diffusion can be arrived at in a similar way through proper derivation of the friction coefficient. Equation (4.70) does not take into account hydrodynamic interactions. It is also necessary to come up with an equivalent radius, r, for a polymer chain, which can be difficult, especially when the conformation is such that the chain is extended, and does not form a sphere at all. Nonetheless, a radius of gyration, rg , is often used to characterize polymer chains in solution, and the resulting friction coefficient is ζ = 4πµrg . More rigorous treatments include hydrodynamic interaction effects and relate the diffusivity directly to the molecular weight of the polymer chain. The general result is that the diffusivity of polymer chains in solution is related to the molecular weight of the polymer, M: DAB ∝ M −ν (4.85) The value of ν can vary. Theory predicts that the diffusivity of polymer chains in solution should be proportional to the inverse square root of the number of beads,

364

TRANSPORT PROPERTIES OF MATERIALS

N −1/2 , and since N is proportional to M, we have DAB ∝ M −1/2 . In “good solvents,” or solvents in which interactions between the polymer chain and solvent are favorable, and the chain takes on an extended structure, we have ν = 3/5. If hydrodynamic interactions are ignored, we expect DAB ∝ 1/M, and ν = 1.0. 4.3.3.2 Diffusion in Polymer Melts. We expect the diffusion situation to be more complex in polymer melts, since chains become entangled and it is difficult for them to move past one another. We further expect that chain entanglements should increase with the molecular weight of the polymer chain. A model has been developed that predicts these molecular weight effects in polymer melts. The reptation model assumes that an isolated polymer chain is contained in a hypothetical tube, as shown schematically in Figure 4.53. Under the constraints imposed by the tube, two types of chain motion are possible. First, there is a conformational change of the chain taking place within the tube, along with a “reptation” motion that causes the chain to translate through the tube, much like a snake wriggling through the grass. Eventually, this motion will carry the chain completely out of the hypothetical tube along what is termed the “primitive path.” The time-dependent evolution of the primitive path leads to chain disentanglement, and it is the basis for determination of the diffusion coefficient. The diffusivity in polymer melts is predicted by this model to vary inversely with the square of molecular weight—that is, ν = 2 in Eq. (4.85). This dependence has been observed experimentally, as shown in Figure 4.54 for a series of deuterated polyethylene (PE) chains in a PE melt. A non-entangled polymer melt is expected to follow the model of non-interacting spheres, and ν = 1.0. There can be a transition from non-entangled to entangled melts, depending on time and conditions. 4.3.3.3 Permeability and Partition Coefficients. In the solid form, many polymers are used in packaging. The diffusion of gases through the packaging material can be important in a number of applications, especially where food and beverages are concerned. Polymeric films can also be used as membranes, such as in reverse osmosis applications, in fuel cells, or as replacements for biological materials in artificial kidneys, intestines, or other organs. Membranes are thin layers of materials used to control access of gases, liquids, or solids in solutions. We briefly describe here the

Figure 4.53 Schematic illustration of the reptation process in polymer melts, showing chain entanglements (light arrows), the “wriggling motion” of the polymer chain (darker arrows), and the primitive path of the polymer chain (dark line). Reprinted, by permission, from G. Strobl, The Physics of Polymers, 2nd ed., p. 283. Copyright  1997 by Springer-Verlag.

MASS TRANSPORT PROPERTIES OF MATERIALS

365

D [m2s−1]

10−12

10−13

10−14

104 M

Figure 4.54 Diffusivity at 176◦ C for polyethylene of different molecular weights, M. The continuous line corresponds to DAB ∝ M −2 . Reprinted, by permission, from G. Strobl, The Physics of Polymers, 2nd ed., p. 286. Copyright  1997 by Springer-Verlag.

transport of molecules through nonporous polymeric films, primarily for membrane applications. There are also porous membranes, which allow particles smaller than the pore diameter to pass through while retaining the larger particles, but the principles of diffusion in porous solids of the previous section are applicable there. In nonporous membranes, diffusion occurs as it would in any other nonporous solid. However, the molecular species must first dissolve into the membrane material. This step can oftentimes be slower than the diffusion, such that it is the rate-limiting step in the process. As a result, membranes are not characterized solely in terms of diffusion coefficients, but in terms of how effective they are in promoting or limiting both solubilization and diffusion of certain molecular species or solutes. When the solute dissolves in the membrane material, there is usually a concentration discontinuity at the interface between the membrane and the surrounding medium (see Figure 4.55). The equilibrium ratio of the solute concentration in one medium, c1 , to the solute concentration in the surrounding medium, c2 , is called the partition coefficient, K12 , and can be expressed in terms of either side of the membrane. For the water–membrane–water example illustrated in Figure 4.55, K12 =

cm2 cm1 = cw1 cw2

(4.86)

The partition coefficient is thus dimensionless. Values for partition coefficients vary greatly, depending on the solute and the membrane material. The concept of permeability, PM , described first in Section 4.3.2.2 also applies to membranes. Equation (4.77) relates the permeability to the diffusion coefficient and solubility. Some representative values of permeabilities for common gases in common polymer films are given in Table 4.17. The units of permeability in Table 4.17 are obtained when diffusivity is in units of m2 /s, and gas solubility is in units of m3 gas·m2 /(m3 solid·N). Note that carbon dioxide permeabilities are generally 3–4 times

366

TRANSPORT PROPERTIES OF MATERIALS

WATER

MEMBRANE

WATER

Cw1

Cm1

Cw2 Cm2

Figure 4.55 Schematic illustration of concentration drop at a membrane interface, and concentration gradient across the membrane. From A. T. Johnson, Biological Process Engineering. Copyright  1999 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Table 4.17 Permeabilities of Selected Polymers Permeability [1016 m4 /(s N)] Film Type

Oxygen

Carbon Dioxide

Water Vapor

Low-density polyethylene Linear low-density polyethylene Medium-density polyethylene High-density polyethylene Polypropylene Polyvinyl chloride Polyvinyl chloride, plasticized Polystyrene Ethylene vinyl acetate copolymer (12%) Ionomer Ruber hydrochloride Polyvinylidene chloride

0.45–1.5 0.80–1.1 0.30–0.95 0.059–0.46 0.15–0.73 0.071–0.26 0.0088–0.086 0.023–0.088 0.091–1.5 0.40–0.86 0.015–0.15 0.00091–0.0030

0.88–8.8 — 0.88–4.4 0.45–1.1 0.88–2.4 0.49–0.93 0.088–6.3 1.1–3.0 4.0–6.1 1.1–2.0 0.059–0.59 0.0067

0.97–3.8 2.6–5.0 1.3–2.4 0.65–1.6 0.65–1.7 — >1.3 18–25 9.7 3.6–4.9 >1.3 0.24–0.81

Source: A. T. Johnson, Biological Process Engineering. Copyright  1999 By John Wiley & Sons, Inc.

MASS TRANSPORT PROPERTIES OF MATERIALS

367

higher than for oxygen, due primarily to the higher solubility of carbon dioxide in most hydrocarbon materials. To obtain the volume flow rate of gases through these films, a modified form of Fick’s Law may be used PM A P (4.87) V˙ = L

Cooperative Learning Exercise 4.12

A 0.15-mm-thick film is required for use in packaging a pharmaceutical product at 30◦ C. The partial pressure of O2 is 0.21 atm outside the package and is 0.01 atm inside the package. Use permeability data from Table 4.17, and assume that the resistance to diffusion outside and inside the membrane are negligible compared to the resistance of the film. Person 1 : Calculate the volume flow rate per unit area of oxygen through a linear low-density polyethylene film. Person 2 : Calculate the volume flow rate per unit area of oxygen through a low-density polystyrene film. Compare your answers. Which is better for this type of application? Which do you think would be preferable under high relative humidity, assuming that water vapor could potentially damage the product? Answers: (Average values of PM used) V˙ /A = 1.3 × 10−8 m3 /s·m2 for polyethylene V˙ /A = 7.5 × 10−10 m3 /s·m2 for polystyrene PS is better for this application (lower flow rate of oxygen across barrier), but its permeability for water is higher than PE.

where V˙ is the volumetric flow rate of gas through the film in m3 /s, PM is the permeability, A is the membrane cross-sectional area in m2 , L is the membrane thickness in meters, and P is the partial pressure difference of the diffusing gas across the membrane in Pa. The permeability is affected not only by temperature, but also by relative humidity. For example, cellophane is an excellent oxygen blocker when dry, but becomes poor when moist. The values for permeability in Table 4.17 are strictly for flat materials. Real permeabilities can vary greatly due to flexural cracks, overlapping, and delamination that are common in real materials. There have been attempts, however, to model the diffusion in strained materials [16]. 4.3.4

Mass Transport Properties of Composites

Little is known about the mass transport properties of reinforced-composite materials. Certainly, there are no new relations or concepts that govern estimations of diffusivities that have not already been discussed. In most polymer–matrix composites, the transport properties of the polymer play an important role in diffusion through the composite. For example, hydrophilic polymers such as epoxy readily absorb water from the atmosphere. Thermoplastic polymers absorb relatively little moisture since they are more hydrophobic, but are more susceptible to uptake of organic solvents.

368

TRANSPORT PROPERTIES OF MATERIALS

Structural effects can also be very important in diffusion through composites. For example, when pores are created either during formation or through de-bonding between phases, diffusion in porous solids as described in Section 4.3.2.3 can occur. More importantly, atomic mobility can occur along the matrix–reinforcement interphase and can lead to composite degradation. Not only is diffusion important in this case, but such transport processes as wicking and capillary action can occur, especially when liquids are involved. For this reason, environmental barrier coatings (EBCs) are often employed, which provide only an outer layer of protection, in this case a diffusion barrier, but contribute nothing from a structural standpoint. Recall from Chapter 1 that this is an example a complementary composite effect. Environmental barrier coatings are a type of laminar composite. As with heat transfer, diffusion in laminar composites can be modeled as steady state diffusion through a composite wall, as illustrated in Figure 4.56. Here, hydrogen gas is in contact with solid material A at pressure P1 and in contact with solid B at pressure P2 . At steady state, ∗ ∗ the molar flux of hydrogen through both walls must be the same (i.e., JH,Ax = JH,Bx ), and Fick’s Law [Eq. (4.4)] in the x direction becomes DH A

c ∗ − c2 c1 − c1∗ = DH B 2 x1 x2

(4.88)

The concentrations in the solid phases, c1∗ and c2∗ , are determined by the solubilities and diffusivities of hydrogen in A and B, and so they are not equal. The thermodynamic activity of hydrogen has a single value at the interface, however. (Refer to Section 3.0.1 for a description of thermodynamic activity.) Hence, the treatment of diffusion flux in a composite wall is simplified by considering activity gradients rather than concentration gradients. If the dissolution of hydrogen gas in the solid follows the reaction H (dissolved )

H2 gas at pressure P1

A

(4.89)

B

C1 C *2 C *1

C2

JH (in A)

JH (in B )

∆x

∆x

H2 gas at pressure P2

1 H (g) 2 2

Figure 4.56 Schematic representation of concentration profiles of hydrogen diffusing through a composite wall. Reprinted, by permission, from D. R. Gaskell, An Introduction to Transport Phenomena in Materials Engineering, p. 498. Copyright  1992 by Macmillan Publishing.

MASS TRANSPORT PROPERTIES OF MATERIALS

369

then the Law of Mass Action [Eq. (3.4)] tells us that the equilibrium constant for the reaction, K, is given by [H (dissolved )] (4.90) K=  1/2 H2 (g) Recall that the concentration (and activity) of molecular hydrogen in the gas phase can be represented by its partial pressure, PH2 , and the concentration of dissolved hydrogen is represented as cH (either c1 or c2 ) in Figure 4.56, so that the equilibrium constant becomes cH K = 1/2 (4.91) PH2

There is an equilibrium constant for the dissolution of hydrogen in each solid phase, KA and KB , respectively. Similarly, the activity of hydrogen at the interface, aH∗ is related to the concentration at the interface and equilibrium constants, c1∗ = KA aH∗ and c2∗ = KB aH∗ , so that Eq. (4.88) becomes 1/2

DH A KA

P1

1/2

− aH∗ a ∗ − P2 = DH B KB H x1 x2

(4.92)

and we can solve for the activity of the interface to obtain 1/2

aH∗ =

4.3.5

1/2

DHA KA x2 P1 + DHB KB x1 P2 DHA KA x2 + DHB KB x1

(4.93)

Mass Transport Properties of Biologics

As was the case for composites, there is little new in the way of fundamental concepts for mass transport in biologics that has not already been presented. However, it is possible to briefly describe extensions of some previously introduced topics that are of particular importance to biological materials—namely, diffusion of nonspherical molecules in solution, diffusion through biological membranes, and convective mass transfer in biological systems. 4.3.5.1 Diffusion in Biological Fluids. Recall from Section 4.3.3.1 that polymer molecules in dilute solutions can be characterized by a friction coefficient, ζ , which describes the resistance to motion through the solvent, and that the diffusivity of these molecules can be related to the friction coefficient by

DAB =

kB T ζ

(4.84)

Recall also from Section 4.1.5 that many biological molecules in solution are not spherical, and can be modeled as either prolate or oblate ellipsoids (see Figure 4.17). A further factor affecting diffusion of these ellipsoids in dilute aqueous solutions is that the molecules can become solvated, resulting in an effective radius that is different than the unsolvated molecules. These concepts of asymmetry and solvation are illustrated in Figure 4.57. Accordingly, there are friction coefficients for the unsolvated Stoke’s

370

TRANSPORT PROPERTIES OF MATERIALS

Cooperative Learning Exercise 4.13 Consider the diffusion of hydrogen through a composite wall comprising Ni and Pd of identical thicknesses at 400◦ C. The hydrogen gas pressure in contact with the Ni is 1.0 atm, and that in contact with the Pd is 0.1 atm. The following data on each metal are provided:

Atomic weight, kg/kg-mol Density at 400◦ C, kg/m3 Diffusivity of H at 400◦ C, m2 /s Solubility of H at 400◦ C and 1 atm hydrogen pressure, atomic ppm

Ni

Pd

58.71 8770 4.9 × 10−10 250

106.4 11,850 6.4 × 10−9 10,000

Person 1 : Calculate the equilibrium constant for the dissolution of hydrogen in nickel, KNi . Person 2 : Calculate the equilibrium constant for the dissolution of hydrogen in palladium, KPd . (Hint: Determine concentration in each metal by converting solubility to weight and volume.) Combine your answers to calculate the activity of hydrogen at the interface. Person 1 : Use the activity of hydrogen at the interface to calculate the concentration of hydrogen in nickel at the interface. Person 2 : Use the activity of hydrogen at the interface to calculate the concentration of hydrogen in palladium at the interface. Combine your information to calculate the flux of hydrogen (in kg/m2 ·s) through the composite wall if both metals have a thickness of 1 mm. Answers: KNi = 0.0374/(1.0)0.5 = 0.0374; KPd = 1.11/(1.0)0.5 = 1.11; a ∗ H = 0.318 c∗ 1 = 0.0374(0.318) = 0.0119 kg/m3 ; c∗ 2 = 1.11(0.318) = 0.353 kg/m3 J ∗ H−Ni = 4.9 × 10−10 (0.0374 − 0.0119)/0.001 = 1.25 × 10−8 kg/m2 ·s (Adapted from Gaskell)

sphere, ζ0 , the actual particle, ζ , and the equivalent solvated spheres, ζ ∗ . The ratio of the friction coefficients, ζ /ζ ∗ , then is a measure in the increase in ζ due to particle asymmetry, and the ratio ζ ∗ /ζ0 is a measure of the increase in friction coefficient due to solvation. The first ratio can be related to the axial ratios (b/a) of the ellipsoids. For prolate ellipsoids (b/a < 1) ζ = ζ∗

[1 − (b/a)2 ]1/2   1 + [1 − (b/a)2 ]1/2 2/3 (b/a) ln b/a

(4.94)

and for oblate ellipsoids (b/a > 1) [(b/a)2 − 1]1/2 ζ = ζ∗ (b/a)2/3 tan−1 [(b/a)2 − 1]1/2

(4.95)

MASS TRANSPORT PROPERTIES OF MATERIALS

371

R

Ro Unsolvated sphere

Solvated sphere

R

Sphere excluded by tumbling ellipsoid

Figure 4.57 Schematic illustration of effective radius of solvated ellipsoids. Reprinted, by permission, from P. Hiemenz, Polymer Chemistry: The Basic Concepts, p. 626. Copyright  1984 by Marcel Dekker, Inc.

The second ratio, ζ ∗ /ζ0 , is related to the density and mass of the solvated and unsolvated molecules:   m1b ρ2 1/3 ζ∗ = 1+ (4.96) ζ0 m2 ρ 1 where m1b is the mass of bound solvent, m2 is the mass of the solute molecule, and ρ1 and ρ2 are the density of the solvent and solute, respectively. Note that multiplication of the two ratios gives the third ratio, ζ /ζ0 . A graphical summary of the variation in this ratio with axial ratio and concentration for aqueous protein dispersions is given in Figure 4.58. With Eqs. (4.94)–(4.96) and Eq. (4.84), it is possible to estimate the diffusivity for solvated ellipsoids. 4.3.5.2 Diffusion through Biological Membranes. Diffusion through polymeric membranes has already been described, and while there are many examples of membrane diffusion in biological systems, such as dialysis, reverse osmosis, and cell metabolism, the principles described previously are still applicable to biological materials. We take a moment here to elaborate on only one example topic, namely, the role of skin as a biological membrane. In a way, skin is a type of environmental barrier coating (cf. Section 4.3.4). Its purpose is to separate an organism from its environment and to provide a barrier to certain substances. Fortunately (and unfortunately), skin is not totally impervious to all substances, particularly air and water. Some animals can obtain a reasonably large fraction of their oxygen requirements through the skin. For example, humans obtain about 2% of the resting oxygen demand through the skin. Passive diffusion of water vapor through the skin is about 6% of the maximum sweat capacity in humans. Water can also be absorbed through the skin, where the mass diffusivity is about 2.5 to 25 × 10−12 m2 /s.

372

TRANSPORT PROPERTIES OF MATERIALS

20.0 10.0 6.0 z z0 = 1.70

4.0 3.0

1.65

1.60

1.55

1.50

1.45

1.40

1.35

1.30

1.25

1.20

1.15

1.10

1.0

1.05

a b

2.0

0.6 0.4 0.3 0.2

0.1

0.06

0.4

0.8

0.1 1.6 grams water gram protein

2.0

2.4

2.8

Figure 4.58 Effect of axial ratio, a/b, and concentration on friction coefficient ratio for aqueous protein dispersions. Reprinted, by permission, from P. Hiemenz, Principles of Colloid and Surface Chemistry, 2nd ed., p. 88. Copyright  1986 by Marcel Dekker, Inc.

Cooperative Learning Exercise 4.14

The diffusivity of human hemoglobin at 20◦ C is 6.9 × 10−11 m2 /s. The viscosity of water at 20◦ C is approximately 0.01 poise, and the radius of a hemoglobin molecule in solution is approximately 2.64 × 10−9 m. Person 1 : Estimate the experimental friction coefficient, ζ , from the experimental diffusivity for human hemoglobin using Equation 4.84. Person 2 : Estimate the unsolvated Stoke’s sphere friction coefficient, ζ0 , using the Stokes–Einstein relationship for spheres, ζ0 = 6πµr. Combine your answers to obtain the ratio of friction coefficients, ζ /ζ0 . If the axial ratio of hemoglobin is a/b = 1.2, use Figure 4.58 to estimate the extent of hydration for hemoglobin in solution. Answers: ζ = (1.38 × 10−23 )(293)/6.9 × 10−11 ) = 5.86 × 10−11 kg/s; ζ0 = 6π(10−3 )(2.64 × 10−9 ) = 4.98 × 10−11 kg/s ζ /ζ0 = 5.86/4.98 = 1.18; from Figure 4.58, fraction solvation is about 0.5 grams water/gram hemoglobin.

MASS TRANSPORT PROPERTIES OF MATERIALS

373

In a manner similar to a laminar composite material, skin is composed of two major layers: the nonvascular epidermis and the highly vascularized (blood-vesselcontaining) dermis. The outer layer of the epidermis is the stratum corneum, which is about 10 µm thick and is composed of compacted dead skin cells. This is the greatest barrier to permeability of penetrating substances. The stratum corneum provides the greatest barrier against hydrophilic compounds, while the viable epidermis and dermis are most resistant to lipophilic compounds. Attempts have been made to increase the permeability of the skin, as in the case of transdermal drug delivery. For example, the diffusivity of the stratum corneum to certain compounds can increase as much as tenfold when the layer is hydrated. Hydrating agents such as dimethyl sulfoxide (DMSO) and surfactants may be added to patches to enhance drug movement through the stratum corneum. Some values of permeability to various commercial solvents of female breast epidermis (300–600 µm thick) are given in Table 4.18. Note that these units of permeability are different than those in Table 4.17 for polymers. Units of m/s can be obtained by multiplying values in Table 4.18 by the appropriate solvent liquid density. 4.3.5.3 Convective Mass Transfer in Biological Systems*. In a manner analogous to convective heat transfer (cf. Section 4.2.5.2), mass transfer can occur not only by static diffusion, but by convection as well. Convective mass transfer occurs when mass is moved in bulk, either as a component in a moving medium or by itself due to pressure or surface tension gradients. As in the case with convective heat transfer, we briefly describe here the dimensionless groups that are involved in convective mass transfer, not as a thorough analysis of the topic, but rather to provide a means for estimating important physical parameters for the materials of interest—specifically, diffusivity. Although the focus here is on convective mass transfer in biologics, primarily due to some important processes such as drying and membrane separations that Table 4.18 Skin Permeability to Various Commercial Solvents

Solvent Dimethyl sulfoxide (DMSO) N-Methyl-2-pyrrolidone Dimethyl acetamide Dimethyl formamide Methyl ethyl ketone Methylene chloride Water Ethanol Butyl acetate Gammabutyrolactone Toluene Propylene carbonate Sulfolane

Permeability [10−7 kg/(m2 s)] 489 475 297 272 147 66.7 41.1 31.4 4.4 3.1 2.2 1.9 0.6

Source: A. T. Johnson, Biological Process Engineering. Copyright  1999 by John Wiley & Sons, Inc.

374

TRANSPORT PROPERTIES OF MATERIALS

are important in biological processes, the discussion here is general to all materials classes. The mass transfer analog to the Prandtl number, Pr, is the Schmidt number, Sc: Sc =

µ ρDAB

(4.97)

where µ is the fluid viscosity, ρ the fluid density, and DAB the mass diffusivity of dilute solute through the fluid. The Schmidt number is the ratio of momentum diffusivity to mass diffusivity, just as the Prandtl number is the ratio of momentum diffusivity to thermal diffusivity. The Nusselt number [Eq. (4.64)] analog in convective mass transport is the Sherwood number, Sh: Sh =

kx l cDAB

(4.98)

where kx (in molar units) is the mass transfer analog to the heat transfer coefficient, hc , [see Eq. (4.94)], l is a characteristic length, c is the concentration (in molar units), and DAB is again the diffusivity. The characteristic length, l, may be a distance along a straight path of mass transfer, or it may be a diameter or radius for mass transfer to and from circular or spherical geometries. Just as there are correlations for the Nusselt number based on the Reynolds and Prandtl numbers [Eq. (4.96)], there are empirical correlations for the Sherwood number as a function of Reynolds and Schmidt numbers, which are of the general form Sh = constant Ren Scm

(4.98)

The values of n and m vary depending on the system and the geometry, but typical values are n = m = 0.5 for creeping flow around a sphere in a gas–liquid system, and n = m = 0.33 for creeping flow around a sphere in liquid–solid systems. REFERENCES Cited References 1. Cowie, J. M. G., Polymers: Chemistry and Physics of Modern Materials, 2nd ed., p. 252, Chapman and Hall, New York, 1991. 2. Oommen, Z., S. Thomas, C. K. Premalatha, and B. Kuriakose, Melt rheological behavior of natural rubber/poly(methyl methacrylate)/natural rubber-g-poly(methy methacrylate) blends, Polymer, 38(22), 5611–5621 (1997). 3. Sood, R., M. Kulkarni, A. Dutta, and R. A. Mashelkar, Polym. Eng. Sci., 28, 20 (1988). 4. Caneiro, O. S., and J. M. Maia, Rheological behavior of (short) carbon fiber/thermoplastic composites. Part II: The influence of matrix type, Polym. Composites, 21(6), 970 (2000). 5. Green, J. P., and J. O. Wilkes Steady State and Dynamic Properties of Concentrated FiberFilled Thermoplastics, Pol. Eng. Sci., 32(21), 1670 (1995). 6. Shuler, S. F., D. M. Binding, and R. B. Pipes, Rheological behavior of two- and three-phase fiber suspensions, Polym. Composites, 15(6), 427 (1994). 7. Jamil, F. A., M. S. Hameed, and F. A. Stephan, Rheological, mechanical, and thermal properties of glass-reinforced polyethylenes, Polym.-Plast. Technol. Eng., 33(6), 659 (1994).

REFERENCES

375

8. Suhara, F., and S. K. N. Kutty, Rheological properties of short polyester fiber-polyurethane elastomer composite with different interfacial bonding agents, Polym. Plast. Technol. Eng., 37(1), 57 (1998). 9. Two Phase Polymer Systems, L. A. Utracki, ed., Hanser, Munich, 1991, p. 305. 10. Castro, J. M., and C. W. Macosko, AIChE J., 28, 250 (1982). 11. Bird, R. B., W. E. Stewart, and E. N. Lightfoot, Transport Phenomena, 2nd ed., John Wiley & Sons, New York, 2001, p. 279. 12. Ammar, M. M., et al., Thermal conductivity of silicate and borate glasses, J. Am. Ceram. Soc., 66, C76–C77 (1983). 13. Primenko, V. I., Theoretical method of determining the temperature dependence of the thermal conductivity of glasses, Glass Ceram., 37, 240–242 (1980). 14. Scholze, H., Glass: Nature, Structure and Properties, Springer-Verlag, New York, 1991, p. 364. 15. Bird, R. B., W. E. Stewart, and E. N. Lightfoot, Transport Phenomena, 2nd ed., John Wiley & Sons, 2001, p. 282. 16. Hinestroza, J., D. De Kee, and P. Pintauro, Apparatus for studying the effect of mechanical deformation on the permeation of organics through polymer films, Ind. Eng. Chem. Res., 40, 2183 (2001).

Transport Phenomena Bird, R. B., W. E. Stewart, and E. N. Lightfoot, Transport Phenomena, 2nd ed., John Wiley & Sons, New York, 2001. Hirschfelder, J. O., C. F. Curtiss, and R. B. Bird, Molecular Theory of Gases and Liquids, John Wiley & Sons, New York, 1954. Gaskell, D. R., An Introduction to Transport Phenomena in Materials Engineering, Macmillan, New York, 1992.

Transport Properties of Metals and Alloys Iida, T., and R. I. L. Guthrie, The Physical Properties of Liquid Metals, Clarendon Press, Oxford, 1988.

Transport Properties of Ceramics and Glasses Scholze, H., Glass: Nature, Structure and Properties, Springer-Verlag, New York, 1991. Experimental Techniques of Glass Science, C. J. Simmons and O. H. El-Bayoumi, ed., American Ceramic Society, Westerville, OH, 1993.

Transport Properties of Polymers Polymer Blends and Alloys, M. J. Folkes and P. S. Hope, eds., Blackie Academic & Professional, New York, 1993. Two-Phase Polymer Systems, L. A. Utracki, ed., Hanser, New York, 1991. Polymer Blends and Composites, J. A. Manson and L. H. Sperling, Plenum, New York, 1976. Dynamics of Polymeric Liquids, Bird, R. B., R. C. Armstrong, and O. Hassager, Vol. 1, John Wiley & Sons, New York, 1977.

Transport Properties of Composites Composites Engineering Handbook, P. K. Mallick, editor, Marcel Dekker, New York, 1997.

376

TRANSPORT PROPERTIES OF MATERIALS

Transport Properties of Biologics Johnson, A. T., Biological Process Engineering, Wiley-Interscience, New York, 1999. Lightfoot, E. N., Transport Phenomena and Living Systems, John Wiley & Sons, New York, 1974.

PROBLEMS Level I

4.I.1 Estimate the viscosity of liquid chromium at its melting temperature of 2171 K. 4.I.2 Estimate the viscosity of liquid iron at 1600◦ C. The density of liquid iron at 1600◦ C is 7160 kg/m3 . 4.I.3 Various molecular weight fractions of cellulose nitrate were dissolved in acetone and the intrinsic viscosity was measured at 25◦ C: M × 10−3 (g/mol) 77 89 273 360 400 640 846 1550 2510 2640 [η] (dl/g) 1.23 1.45 3.54 5.50 6.50 10.6 14.9 30.3 31.0 36.3 Use these data to evaluate the Mark–Houwink constants for this system. 4.I.4 Agglomeration in a slurry causes a change in fcrν from 0.6 to 0.45 and a change in KH from 3.5 to 2.5. Calculate and compare the viscosity for each slurry when fpν = 0.40. 4.I.5 Calculate the thermal resistance of a 2-mm-thick section of human skin. The thermal conductivity of human skin is 0.627 N/K·s, the specific heat is 3470 N·m/kg·K, and the density is 1100 kg/m3 . 4.I.6 A plate of iron is exposed to a carburizing (carbon-rich) atmosphere on one side and a decarburizing (carbon-deficient) atmosphere on the other side at 700◦ C. If a condition of steady state is achieved, calculate the diffusion flux of carbon through the plate if the concentration of carbon at positions of 5 and 10 mm beneath the carburizing surface are 1.2 and 0.8 kg/m3 , respectively. Assume a diffusion coefficient of 3 × 10−11 m2 /s at this temperature. Level II

4.II.1 The thermal conductivity of a material varies with temperature as ln k = 0.01T + 0.05 where k has units of W/m·K, and T has units of ◦ C. Heat flows by conduction through a plane slab of this material of thickness 0.1 m, the left face of which is at 100◦ C and the right face of which is at 0◦ C. Calculate (a) the mean thermal conductivity of this material in the range 0–100◦ C, (b) the heat flux through the slab, and (c) the temperature at x = 0.05 m.

PROBLEMS

377

4.II.2 Hydrogen dissolved in iron obeys Sievert’s Law ; that is, at equilibrium we have 1 H (g) 2 2

[H]dissolved

in iron

where [H] is the concentration of atomic hydrogen in the iron, in ppm by weight, in equilibrium with hydrogen gas at the pressure PH2 . At 400◦ C and a hydrogen pressure of 1.013 × 105 Pa (1 atm), the solubility of hydrogen in iron is 3 ppm by weight. Calculate the flux of hydrogen through an iron tank of wall thickness 0.001 m at 400◦ C. The density of iron 400◦ C is 7730 kg/m3 , and the diffusivity of hydrogen is 10−8 m2 /s. Clearly state any assumptions you make. 4.II.3 A hemodialysis membrane with an effective area of 0.06 m2 , thickness of 50 µm, and permeability of 2.96 × 10−14 m2 is used to filter urea and other impurities from the blood. The viscosity of blood plasma is 1.2 × 10−3 N·s/m2 . What is the expected filtration rate if the transmembrane pressure is 2.25 Pa? 4.II.4 Fick’s Second Law of Diffusion relates the change in concentration of a diffusing species with time to the diffusion coefficient and the concentration gradient for non-steady-state diffusion: ∂cA ∂ 2 cA =D 2 ∂t ∂x The solution to this partial differential equation depends upon geometry, which imposes certain boundary conditions. Look up the solution to this equation for a semi-infinite solid in which the surface concentration is held constant, and the diffusion coefficient is assumed to be constant. The solution should contain the error function. Report the following: the boundary conditions, the resulting equation, and a table of the error function. Level III

4.III.1 A relationship for the viscosity of a simple liquid metal near its melting point, Tm , was proposed by Andrade: −4

µ = 1.68 × 10

√ MTm 2/3

Vm

where µ is in mPa·s and Vm is the atomic volume at the melting point. Show that this relationship gives approximately the same value for viscosity (within a factor of about 0.3) as the hard-sphere model when a packing fraction of 0.45 is used. Use this formula to estimate the viscosity of molten silver near its melting point. How does this compare to an experimental value? Cite your references. 4.III.2 Use matrix manipulations to solve for the three constants in the Vogel– Fulcher–Tammann using the data in Cooperative Learning Exercise 4.1. Show your work.

378

4.III.3

TRANSPORT PROPERTIES OF MATERIALS

Determine the VFT equation for a pure SiO2 melt based on the following three viscosity–temperature points: Glass transformation temperature = 1203◦ C Littleton temperature = 1740◦ C Working point = 2565◦ C

ln µ = 3.3 ln µ = 7.6 ln µ = 4.0

4.III.4

In Cooperative Learning Exercise 4.4, you derived an expression for the superficial velocity of a 1% solution of poly(ethylene oxide) through a bed of 10-µm-diameter fibers with 60% porosity. You were left with a parameter, τRh , in the expression, which is the shear stress at the wall, based upon the hydrodynamic radius, Rh . The mean hydrodynamic radius is defined as the ratio of the particle cross section to the “wetted perimeter,” which is conveniently related to the void fraction, or porosity, ε, and the wetted surface, a, as Rh = ε/a. The wetted surface, in turn, is related to the porosity and the “specific surface,” S, as a = S(1 − ε). Finally, the specific surface is directly related to geometrical dimensions of the particle and is defined as the ratio of particle surface area to particle volume. Complete the derivation of an expression for the superficial velocity by calculating the specific surface for cylinders (this result is general for all cylindrical geometries) and calculating a value of τRh for the case of 60% porosity. Your final expression for v0 should have only the pressure drop across the fiber bed, P , and the bed length, L.

4.III.5

Derive an expression for steady-state heat flow through three parallel planes as shown below. In this instance, the total heat flow, Qtot , is a sum of the individual heat flows, Qi . The cross-sectional areas of each plane, Ai , are not necessarily equivalent. T1

T2

QA

QB

QC

∆x

4.III.6

Composite membranes are used where the membrane of choice for a particular separation must be strengthened by addition of a second material. Assume

PROBLEMS

379

a hollow fiber membrane of inside diameter 1.2 mm and outside diameter 1.3 mm. Mass diffusivity of the target species through this membrane is 3.0 × 10−10 m2 / s. A second material is layered on the outside of the first material to add strength. Mass diffusivity of the same target species through this material is 18 × 10−10 m2 / s. What thickness of the outside material can be added if the total diffusion does not exceed the resistance of the inner material by more than 25%?

CHAPTER 5

Mechanics of Materials

5.0

INTRODUCTION AND OBJECTIVES

The engineering principles of thermodynamics, kinetics, and transport phenomena, as well as the chemistry and physics of molecular structure, serve as the basis for the remaining topics in this book. In this chapter, we look at what for many applications is the primary materials selection criterion: mechanical properties. As in the previous chapter, we focus primarily on the properties of materials, but discuss briefly the mechanics, both in the fluid and solid states, that give rise to the properties. There is a great deal of new terminology in this chapter, and it is cumulative—take time to understand all the definitions before proceeding to the next section. We seek to understand the response of a material to an applied stress. In Chapter 4, we saw how a fluid responds to a shearing stress through the application of Newton’s Law of Viscosity [Eq. (4.3)]. In this chapter, we examine other types of stresses, such as tensile and compressive, and describe the response of solids (primarily) to these stresses. That response usually takes on one of several forms: elastic, inelastic, viscoelastic, plastic (ductile), fracture, or time-dependent creep. We will see that Newton’s Law will be useful in describing some of these responses and that the concepts of stress (applied force per unit area) and strain (change in dimensions) are universal to these topics. By the end of this chapter, you should be able to: ž ž ž ž ž ž ž ž

Determine strength and modulus from a stress–strain diagram. Qualitatively differentiate between stress–strain diagrams for different material classes. Describe the differences between the Kelvin and Voigt models for viscoelasticity and identify their corresponding equations. Calculate Poisson’s ratio. Relate flexural, bulk, and tensile moduli. Distinguish between ultimate strength, yield strength, and tensile strength. Distinguish between ductile and brittle materials. Understand the role of strain rate on mechanical properties.

An Introduction to Materials Engineering and Science: For Chemical and Materials Engineers, by Brian S. Mitchell ISBN 0-471-43623-2 Copyright  2004 John Wiley & Sons, Inc.

380

MECHANICS OF METALS AND ALLOYS

ž ž

ž

ž

ž

ž

381

Identify the different types of mechanical creep and calculate creep rate. Apply time–temperature superposition principles to polymer moduli and calculate shift factors. Describe temperature, molecular weight, and strain-rate effects on the mechanical properties of polymers. Calculate axial and transverse moduli and strengths for unidirectional, continuous and discontinuous, fiber-reinforced composites and laminate composites. Describe the structure of soft and hard biologics that give rise to their unique mechanical properties. Use mechanical properties to help select appropriate candidate materials for a specific application.

5.1

MECHANICS OF METALS AND ALLOYS

In this section, we will describe the response of a material to an applied stress in terms of its atomic structure. Initially, we will focus on a single crystal of a metallic element, then work our way into topics related to polycrystalline materials. 5.1.1

Elasticity

5.1.1.1 The Molecular Origins of Elasticity. Recall from Section 1.0.4 that atoms are held together by interatomic bonds and that there are equations such as Eq. (1.13) that relate the interatomic force, F , to the potential energy function between the atoms, U , and the separation distance, r:

F =−

dU dr

(1.13)

It should make sense then, that as we attempt to pull atoms apart or force them further together through an applied stress, we can, at least in principle, relate how relatively difficult or easy this is to the potential energy function. We can develop a quantitative description of this process of pulling atoms apart, provided that we do so over small deformations, in which the deformation is wholly recoverable; that is, the atoms can return back to their original, undeformed positions with no permanent displacement relative to one another. This is called an elastic response. The term “elastic” here does not imply anything specific to polymers in the same way that the more everyday use of the term does. It is used in the same sense that it is in physics and chemistry—a completely recoverable deformation. A convenient way to visualize the elastic bond between two atoms is to place an imaginary spring between the atoms, as illustrated in Figure 5.1. For small atomic deformations, either away from each other with what is called a tensile force (shown in Figure 5.1) or toward each other with what is called a compressive force (not shown), the force stored in the spring causes the atoms to return to the undeformed, equilibrium separation distance, r0 , where the forces are zero (cf. Figure 1.3). Attractive forces between the atoms counteract the tensile force, and repulsive forces between the atoms counteract the compressive forces.

382

MECHANICS OF MATERIALS

Example Problem 5.1

Derive an expression for the elastic modulus of an ionic solid in terms of the potential energy function parameters a, b, and n. Solution: According to Eq. (5.5), the elastic modulus, E, is proportional to the stiffness of the theoretical springs that model the bonds between atoms in the solid, S0 . According to Equation (5.1), S0 is in turn proportional to the second derivative of the potential energy function, U , with respect to interatomic separation distance, r. We begin with the potential energy function of Eq. (1.11), and recognize that for an ionic solid, m = 1, such that the potential energy function becomes b −a + n . The strategy, then, is to differentiate U twice with respect to r, and U= r r evaluate at r = r0 . a nb dF −2a dU = 2 − n+1 , and the second derivative is = 3 + The first derivative is F = dr r r dr r n(n + 1)b . All that is left now is to evaluate dF/dr at r = r . Equation (1.16) provides us 0 r n+2 with the equilibrium separation distance as a function of the potential energy parameters,   1 nb n−1 . Substitution of r0 into dF /dr with which for the case of m = 1 becomes r0 = a appropriate simplification gives the desired result: E ∝ S0 =



dU dr



r=r0

=



−2a n(n + 1)b +  3   n+2 nb n−1 nb n−1 a a

F

F r0 r

Figure 5.1 Schematic illustration of “bead-and-spring” model of atomic force between atoms. Reprinted, by permission, from M. F. Ashby and D. R. H. Jones, Engineering Materials 1, 2nd ed., p. 44. Copyright  1996 by Michael F. Ashby and David R. H. Jones.

The stiffness of the spring, S0 , is a measure of how much force must be overcome to pull the atoms apart, and it is related to the potential energy function as S0 =



d 2U dr 2



(5.1) r=r0

Notice that the stiffness is the slope (first derivative) of the force versus interatomic separation distance plot in the vicinity of r0 , as shown in Figure 1.3. So, for small deformations (i.e., r ≈ r0 ) we can relate the stiffness to the applied force, F :

MECHANICS OF METALS AND ALLOYS

S0 =

dF F F F − F0 = ≈ = dr r r − r0 r − r0

383

(5.2)

since F = F0 = 0 at r = r0 (see Figure 1.3). Now consider a square array of N springs (bonds), as would be found in a single crystal, and illustrated in Figure 5.2. The total tensile force exerted across the unit area, F /A, is represented by the Greek lowercase sigma, σt , where the subscript “t” indicates a tensile force. It is the product of the stiffness of each spring, S0 , the number of springs per unit area, N/A, and the distance the springs are extended, r − r0 σt = (N/A)S0 (r − r0 )

(5.3)

The number of springs per unit area is simply 1/r02 , (see Figure 5.2), so that Eq. (5.3) becomes    r − r0 S0 (5.4) σt = r0 r0 We have written Eq. (5.4) with variables grouped as they are in order to define two very important quantities. The first quantity in parentheses is called the modulus—or in this case, the tensile modulus, E, since a tensile force is being applied. The tensile modulus is sometimes called Young’s modulus, elastic modulus, or modulus of elasticity, since it describes the elastic, or recoverable, response to the applied force, as represented by the springs. The second set of parentheses in Eq. (5.4) represents the tensile strain, which is indicated by the Greek lowercase epsilon, εt . The strain is defined as the displacement, r − r0 , relative to the initial position, r0 , so that it is an indication of relative displacement and not absolute displacement. This allows comparisons to be made between tensile test performed at a variety of length scales. Equation (5.4) thus becomes σt = Eεt (5.5)

ro

Unit area, crossed by 1 bonds ro2 s

s

ro ro r

Figure 5.2 Illustration of multiple bead-and-springs in tensile separation of atomic planes. Reprinted, by permission, from M. R. Ashby and D. R. H. Jones, Engineering Materials 1, 2nd ed., p. 59, Copyright  1996 by Michael F. Ashby and David R. H. Jones.

384

MECHANICS OF MATERIALS

Note that the strain is dimensionless, but is often given in units of length/length (e.g., in./in. or cm/cm) to indicate the length scale of the displacement. Modulus then has the same units of stress, which are typically Pa (N/m2 ) or psi (lbf /in.2 ). The stresses required to separate bonds are surprisingly high, and the resulting extensions are vanishingly small, such that the values of modulus are often reported in MPa (106 Pa) and GPa (109 Pa), or kpsi (103 psi, sometimes called “ksi”) and Mpsi (106 psi).

Cooperative Learning Exercise 5.1

Work in groups of three. Look up the modulus of elasticity for each of the following substances in Appendix 7, and compute the tensile stress, σt , that arises when a cylindrical sample of each is strained to 0.1% of their original length. Assume that all deformations are perfectly elastic. Person 1: Tungsten (W) Person 2: Tungsten carbide (WC) Person 3: Polycarbonate Compare your answers. Assuming that all samples have the same geometry (crosssectional area), which of the three will require the greatest load to achieve this amount of strain? Answers: σt (W) = (400 GPa)(0.001) = 400 MPa; σt (WC) = (686)(0.001) = 686 MPa; σt (Polycarbonate) = (2.4 GPa)(0.001) = 2.4 MPa. WC will require the greatest load (force).

Equation (5.5) is known as Hooke’s Law and simply states that in the elastic region, the stress and strain are related through a proportionality constant, E. Note the similarity in form to Newton’s Law of Viscosity [Eq. (4.3)], where the shear stress, τ , is proportional to the strain rate, γ˙ . The primary differences are that we are now describing a solid, not a fluid, the response is to a tensile force, not a shear force, and we do not (yet) consider time dependency in our tensile stress or strain. The relationship between the potential energy and the modulus is an important one. If we know the potential energy function, even qualitatively or relatively to another material, we can make inferences as to the relative magnitude of the elastic modulus. Figure 5.3 shows the interrelationship between the potential energy (top curves), the force (represented per unit area as stress in the bottom curves), and tensile modulus (slope of force curves for two hypothetical materials). The deeper the potential energy well, the larger the modulus (compare Figure 5.3a to 5.3b). 5.1.1.2 Generalized Hooke’s Law∗ . The discussion in the previous section was a simplified one insofar as the relationship between stress and strain was considered in only one direction: along the applied stress. In reality, a stress applied to a volume will have not only the normal forces, or forces perpendicular to the surface to which the force is applied, but also shear stresses in the plane of the surface. Thus there are a total of nine components to the applied stress, one normal and two shear along each of three directions (see Figure 5.4). Recall from the beginning of Chapter 4 that for shear stresses, the first subscript indicates the direction of the applied force (outward normal to the surface), and the second subscript indicates the direction of the resulting stress. Thus, τxy is the shear stress of x-directed force in the y direction. Since this notation for normal forces is somewhat redundant—that is, the x component of an

385

MECHANICS OF METALS AND ALLOYS

U

0

U

∆r r0

∆r r0

0

U0 ~ 20 kJ/mol

U0 ~ 400 kJ/mol

12 1.5

8

0.4 0.3

1.0

6 4

0.5

−0.5

−4

9 s 10 N m2

0

∆r r0

s (106 psi)

9 s 10 N m2

0 −2

20

0.1

2

0

0

∆r r0

−0.1

−20 −0.2

−6 −8 −10 −12

40 Slope = E

0.2

(103 psi)

Slope = E

10

0

1

2

3

−1.0

−0.3

−1.5

−0.4

−40

0

1

2

3

(b)

4

(a)

Figure 5.3 Schematic diagrams showing potential energy (top) and applied stress (bottom) versus linear strain for crystalline solids with (a) strong bonds and (b) weak bonds. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc. z szz tzy tzx tyz

txz txy sxx

tyx

syy

y

x

Figure 5.4 Distribution of shear and normal stresses acting on a volume element. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

386

MECHANICS OF MATERIALS

x-directed force is by definition a normal force and not a shear force—the subscripts are sometimes simplified, and σxx = σx , σyy = σy , and σzz = σz . We will maintain the double subscript notation for completeness for the time being. It is often convenient to consider these nine stress components in the form of a stress tensor, σ , which can be represented by an array of the stress components: 

σxx σ ≡  τyx τzx

τxy σyy τzy

 τxz τyz  σzz

(5.6)

In this array, the stress components in the first row act on a plane perpendicular to the x axis, the stress components in the second row act on a plane perpendicular to the y axis, and the stress components in the third row act on a plane perpendicular to the z axis. The stress matrix of Eq. (5.6) is symmetric, that is, the complementary shear stress components are equivalent—for example, τxy = τyx , and so on. There is a corresponding strain tensor (which is not presented here due to its complexity), and for each stress component there is a corresponding strain component. Hence, there are three normal strains, εxx , εyy , and εzz , and six shear strains, which reduce to three due to matrix symmetry: γxy , γyz , and γxz . There is a linear relationship between stress and strain for each component, giving rise to the following six equations, known as the generalized form of Hooke’s Law: σxx = Q11 εxx + Q12 εyy + Q13 εzz + Q14 γxy + Q15 γxz + Q16 εyz σyy = Q21 εxx + Q22 εyy + Q23 εzz + Q24 γxy + Q25 γxz + Q26 εyz σzz = Q31 εxx + Q32 εyy + Q33 εzz + Q34 γxy + Q35 γxz + Q36 εyz τyz = Q41 εxx + Q42 εyy + Q43 εzz + Q44 γxy + Q45 γxz + Q46 εyz

(5.7)

τxz = Q51 εxx + Q52 εyy + Q53 εzz + Q54 γxy + Q55 γxz + Q56 εyz τxy = Q61 εxx + Q62 εyy + Q63 εzz + Q64 γxy + Q65 γxz + Q66 εyz The elastic coefficients, Qij , can also be summarized in a 6 × 6 array: 

Q11  Q21   .  .. Q16

Q12 Q22 .. . Q26

··· ··· .. . ···

 Q16 Q26   ..  .  Q66

(5.8)

It can be shown that for the cross-terms Q21 = Q12 , Q31 = Q13 , and so on, so that of the initial 36 values, there are only 21 independent elastic constants necessary to completely define an anisotropic volume without any geometrical symmetry (not to be confused with matrix symmetry). The number of independent elastic constants decreases with increasing geometrical symmetry. For example, orthorhombic symmetry has 9 elastic constants, tetragonal 6, hexagonal 5, and cubic only 3. If the body is isotropic, the number of independent moduli can decrease even further, to a limiting

MECHANICS OF METALS AND ALLOYS

387

∆L = L − L0

σt

∆d

d L0

L

st

Figure 5.5 Schematic illustration of tensile strain and corresponding lateral strain. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

number of three independent moduli, corresponding to the three forms of stress: tension, compression, and shear. 5.1.1.3 Moduli and Poisson’s Ratio. Let us examine the elastic deformation of a material on a more macroscopic level. Figure 5.5 shows a typical test material under tension on a scale somewhat larger than that represented in Figure 5.4. A stress, σt is once again applied, resulting in a strain εt , which we define as L/L0 . For small deformations, the stress and strain are related by Eq. (5.5). Notice, however, that in order for the material to (ideally) maintain a constant volume, the diameter, d, must decrease by an increment d = d − d0 . There is no requirement, of course, that volume be maintained, but in most real systems, there is a decrease in the cross section with elongation. This response is characteristic of a material and is given by Poisson’s ratio, ν, which is defined as the ratio of the induced lateral strain to the axial strain:

ν=−

d/d0 L/L0

(5.9)

The minus sign in Eq. (5.9) is to account for the fact that d as defined above is usually negative. Thus, Poisson’s ratio is normally a positive quantity, though there is nothing that prevents it from having a negative value. For constant volume deformations (such as in polymeric elastomers), ν = 0.5, but for most metals, Poisson’s ratio varies between 0.25 and 0.35. Values of Poisson’s ratio for selected materials are presented in Appendix 7. Poisson’s ratio provides a basis for relating the lateral response of a material to the axial response. By definition, the forces that cause lateral deformation in a tensile experiment must be shear forces, since they are normal to the direction of the applied

388

MECHANICS OF MATERIALS

τ

a′

a

g

g 45°

d

b′

b

c

τ

Figure 5.6 Schematic illustration of shear strain and stress. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

stress. In a generalized shear situation, as in Figure 5.6, the shear stress, τ , results in a shear strain, γ , which we must define slightly differently than in the tensile experiment. Here, the shear strain is the amount of shear, aa ′ , divided by the distance of which the shear has occurred, ad, so γ = aa ′ /ad. There is still a proportionality between stress and strain (for small deformations), but now we call the proportionality constant the shear modulus, G: τ = Gγ (5.10) The shear modulus and elastic modulus are related through Poisson’s ratio: G=

E 2(1 + ν)

(5.11)

In addition to the tensile and shear moduli, a compressive modulus, or modulus of compressibility, K, exists to describe the elastic response to compressive stresses (see Figure 5.7). The compressive modulus is also sometimes called the bulk modulus. It is the proportionality constant between the compressive stress, σc , and the bulk strain, represented by the relative change in bulk volume, V /V0 . The compressive modulus and elastic modulus are also related through Poisson’s ratio: E (5.12) K= 3(1 − 2ν) From Eqs. (5.11) and (5.12), we can derive a relationship that relates all three moduli to one another: 9GK E= (5.13) 3K + G Equations (5.11), (5.12), and (5.13) allow us to apply some convenient “rules of thumb.” For most hard materials, v ≈ 1/3, such that K ≈ E and G ≈ (3/8)E. For elastomers, putties, gels, and colloidal systems, the compressive modulus is much higher than the other moduli, and the system is considered “incompressible.” In this case, E ≈ 3G.

MECHANICS OF METALS AND ALLOYS

389

sc

∆L = L0 − L ∆d

d L

L0

sc

Figure 5.7 Schematic illustration of compressive strain and lateral strain. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Cooperative Learning Exercise 5.2 A tensile load of 5600 N is applied to along the long axis (z direction) of cylindrical brass rod that has an initial diameter of 10 mm. The resulting deformation is entirely elastic, and it results in a diameter reduction (x direction) of 2.5 × 10−3 mm. Person 1: Calculate the lateral strain, εx . Person 2: Calculate the axial strain, εz . Assume that the modulus of elasticity for brass is 97 GPa. Combine your information to calculate Poisson’s ratio for this material. Person 2: Use this value of Poisson’s ratio to calculate the compressive modulus, K. Person 1: Use the value of Poisson’s ratio to calculate the shear modulus, G. Use Eq. (5.13) to recalculate E from G and K to confirm your results.

Answers: εx = −(2.5 × 10−3 mm)/(10 mm) = −2.5 × 10−4 ; in order to calculate εz , we must first calculate the applied stress, σ = (5600 N)/[π(5 × 10−3 m)2 ] = 71.3 × 106 Pa, then use Hook’s Law and the elastic modulus to calculate the axial strain, εz = σ/E = (71.3 MPa)/(97000 MPa) = 7.35 × 10−4 ; ν = −εx /εz = −(−2.5)/(7.35) = 0.34. K = 97 GPa/{3[1 − (2)(0.34)]} = 101 GPa; G = 97 GPa/{2[1 + (0.34)]} = 36.2 GPa; E = 9(36.2)(101)/[3(101) + 36.2] = 97.0 GPa. It checks out!

5.1.2

Ductility

To this point, we have limited the discussion to small strains—that is, small deviations from the equilibrium bond distance, such that all imposed deformations are completely recoverable. This is the elastic response region, one that virtually all materials possess (see Figure 5.8). What happens at larger deformations, however, is dependent to some

390

MECHANICS OF MATERIALS

Elastic deformation Plastic deformation

B Yield stress

C

Elastic deformation

B

Plastic deformation

Stress

Stress

A

C

Strain (a)

A

Strain (b)

Figure 5.8 Illustration of (a) ideal elastic deformation followed by ideal plastic deformation and (b) typical elastic and plastic deformation in rigid bodies. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed., Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

extent on the type of material under consideration. Beyond the elastic region, we enter a realm of nonrecoverable mechanical response termed permanent deformation. There are two primary forms of permanent deformation: viscous flow and plastic flow. We will reserve the discussion of viscous flow to the appropriate section on mechanical properties of polymers, since this is where viscous flow primarily occurs. Other types of nonelastic responses to applied forces, such as fracture and creep, will also be delayed until later sections. We concentrate now on a type of nonrecoverable deformation called plastic flow, or, more commonly in the terminology of metals and alloys, ductility. As illustrated in Figure 5.8, the plastic deformation region ideally occurs immediately following the elastic response region, and it results in elongation (strain) without additional stress requirement, hence the flat region represented by segment BC in Figure 5.8a. More practically, there are microstructural phenomena that lead to the behavior shown in Figure 5.8b, namely, an increase in stress with % strain in the plastic deformation region. In the following section, we explore the molecular basis for these observations. 5.1.2.1 The Molecular Origins of Ductility—Slip in Single Crystals. Plastic deformation in single crystals occurs by one of two possible mechanisms: slip or twinning. Slip is the process by which bulk lattice displacement occurs by the motion of crystallographic planes relative to one another (see Figure 5.9a). The surface on which slip takes place is often a plane, known as the slip plane. The direction of motion is known as the slip direction. The visible intersection of a slip plane with the outer surface of the crystal is known as a slip band. Twinning is the process in which atoms subjected to stress rearrange themselves so that one part of the crystal becomes a mirror image of the other part. Every plane of atoms shifts in the same direction by an amount proportional to its distance from the twinning plane (see Figure 5.9b). Twinning usually requires a higher shear stress than does slip, and it occurs primarily in BCC and HCP crystal structures. We will focus primarily on slip to describe the plastic deformation process.

MECHANICS OF METALS AND ALLOYS

391

Twinning plane Slip plane

(b)

(a)

Figure 5.9 Schematic illustration of plastic deformation in single crystals by (a) slip and (b) twinning. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

a A (a)

Shear stress

B

Displacement, x

(b)

Figure 5.10 Schematic illustration of (a) relative shear of two planes of atoms in a strained material and (b) shear stress as a function of relative displacement of the planes from their equilibrium positions. Reprinted, by permission, from C. Kittel, Introduction to Solid State Physics, 2nd ed., p. 517. Copyright  1957 by John Wiley & Sons, Inc.

Based on the concepts of intermolecular forces and shear modulus introduced in the previous section, it is relatively easy to estimate the theoretical stress required to cause slip in a single crystal. We call this the critical shear stress, σcr . Refer to Figure 5.10a, and consider the force required to shear two planes of atoms past each other. In the region of small elastic strains, the stress, τ , is related to the displacement, x, relative to the initial interplanar spacing, d, according to a modified form of Eq. (5.10) for the shear modulus, G: x τ =G (5.14) d As a first approximation, we may represent the stress-displacement relation by a sine function (Figure 5.10b), since when atom A in Figure 5.10a is displaced to the point where it is directly over atom B in the plane below, an unstable equilibrium exists and

392

MECHANICS OF MATERIALS

the stress is zero, so τ =

Ga sin(2πx/a) 2πd

(5.15)

where a is the interatomic spacing in the direction of shear. Note that the slope (represented by the dashed line) of the sine wave in Figure 5.10b is simply the shear modulus, G. For small displacements, x, Eq. (5.15) reduces to Eq. (5.14). The critical shear stress, τcr , at which the lattice becomes unstable and begins to plasticly deform is given by the maximum value of τ in Eq. (5.15), or τcr =

Ga 2πd

(5.16)

If we make the approximation that a ≈ d, then τcr ≈

G G ≈ 2π 6

(5.17)

that is, the critical shear stress to plastic flow should be of the order of 1/6 the shear modulus. In reality, the measured critical shear stress is much smaller than this for most single-crystal metals. Further theoretical considerations reduced this value to G/30, but even this is much larger than the critical shear stress observed in most metals. Clearly, we cannot simply extend the theory of elasticity to large strains to determine shear strength—there are other considerations. The primary consideration we are missing is that of crystal imperfections. Recall from Section 1.1.4 that virtually all crystals contain some concentration of defects. In particular, the presence of dislocations causes the actual critical shear stress to be much smaller than that predicted by Eq. (5.17). Recall also that there are three primary types of dislocations: edge, screw, and mixed. Although all three types of dislocations can propagate through a crystal and result in plastic deformation, we concentrate here on the most common and conceptually most simple of the dislocations, the edge dislocation. The movement of an edge dislocation in a single crystal in response to an applied shear stress is illustrated in Figure 5.11. An edge dislocation can move by slip only in its slip plane, which is the plane containing the line of the dislocation and the Burger’s vector (cf. Figure 1.31). Dislocation densities range from 102 to 103 dislocations/cm2 in the best silicon single crystals to as high as 1011 or 1012 dislocations/cm2 in heavily deformed metal crystals. As we will see later on, heat treatment can reduce the dislocation density in metals by about five orders of magnitude. Beside dislocation density, dislocation orientation is the primary factor in determining the critical shear stress required for plastic deformation. Dislocations do not move with the same degree of ease in all crystallographic directions or in all crystallographic planes. There is usually a preferred direction for slip dislocation movement. The combination of slip direction and slip plane is called the slip system, and it depends on the crystal structure of the metal. The slip plane is usually that plane having the most dense atomic packing (cf. Section 1.1.1.2). In face-centered cubic structures, this plane is the (111) plane, and the slip direction is the [110] direction. Each slip plane may contain more than one possible slip direction, so several slip systems may exist for a particular crystal structure. For FCC, there are a total of 12 possible slip systems: four different (111) planes and three independent [110] directions for each plane. The

MECHANICS OF METALS AND ALLOYS

Shear stress

Shear stress

A

B

393

C

A

D

B

C

D

Slip plane Edge dislocation line (a)

(b) Shear stress

A

B

C

D

Unit step of slip

(c)

Figure 5.11 Schematic illustration edge dislocation motion in response to an applied shear stress, where (a) the extra half-plane is labeled as A (cf. Figure 1.28), (b) the dislocation moves one atomic distance to the right, and (c) a step forms on the crystal surface as the extra half-plane exits the crystal. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 155. Copyright  2000 by John Wiley & Sons, Inc.

possible slip systems for BCC and HCP metals are listed in Table 5.1. Body-centered cubic structures have a high number of slip systems and, as a result, undergo extensive plastic deformation; that is, they are highly ductile. The opposite is true of HCP structures, which have relatively few slip systems. For reference, the corresponding twinning systems are listed in Table 5.2. The final factor in determining the critical shear stress for plastic deformation is how the preferred slip planes are oriented relative to the applied stress. Figure 5.12 illustrates how a test sample may be oriented such that the preferred slip plane is not necessarily along the direction of the applied force, F . Shear forces will still exist, and their magnitude will depend not only on the applied stress, but on orientation. The angle φ in Figure 5.12 represents the angle between the normal to the slip plane and the applied stress direction. The angle λ represents the angle between the slip direction and the applied stress direction. Remember that the slip plane and the slip direction are not the same. It is also important to note that the three vectors in Figure 5.12 are

394

MECHANICS OF MATERIALS

Table 5.1 Slip Systems in FCC, BCC, and HCP Metals Slip Plane

Metals

Slip Direction

Number of Slip Systems

Face-Centered Cubic Cu, Al, Ni, Ag, Au

(111)

[110]

12

Body-Centered Cubic α-Fe, W, Mo α-Fe, W α-Fe, K

(110) (211) (321)

[111] [111] [111]

12 12 24

Hexagonal Close-Packed Cd, Zn, Mg, Ti, Be Ti, Mg, Zr Ti, Mg

(0001) (1010) (1011)

[1120] [1120] [1120]

3 3 6

Table 5.2 Twinning Systems in FCC, BCC, and HCP Metals Crystal Structure

Twinning Planes

Twinning Directions

fcc bcc hcp

(111) (112) (1012)

[112] [111] [1011]

not coplanar in general, that is, φ + λ = 90◦ . The resolved shear stress, τr , then is the component of the applied stress, σt , acting along the particular slip plane of interest and can be determined as τr = σt cos φ cos λ (5.18) As mentioned previously, a number of slip systems can operate simultaneously, but there will be one system that has the orientation which affords it the largest resolved shear stress of all the slip systems in operation. This system will be the one that has the maximum geometric factor, (cos φ cos λ)max , since the applied stress is the same for all slip systems. So, the maximum resolved shear stress, τr,max , is given by: τr,max = σt (cos φ cos λ)max

(5.19)

As the applied stress, σt , increases, the maximum resolved shear stress increases according to Eq. (5.19), finally reaching a critical value, called the critical resolved shear stress, τcr , at which slip along the preferred plane begins and plastic deformation commences. We refer to the applied stress at which plastic deformation commences as

MECHANICS OF METALS AND ALLOYS

395

F

f

A l

N sli orm p a pl l t an o e

ection

Slip dir

F

Figure 5.12 Schematic illustration of the relationship between tensile axis, slip plane, and slip direction used in calculating the resolved shear stress for a single crystal. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 160. Copyright  2000 by John Wiley & Sons, Inc.

the yield stress, σy . The yield stress can be determined from Eq. (5.19) in a form of Schmid’s Law, which states that slip takes place along a given slip plane and direction when the corresponding component of shear stress reaches a critical value: σy =

τcr (cos φ cos λ)max

(5.20)

The minimum stress necessary to introduce yielding occurs when a single crystal is oriented such that φ = λ = 45◦ , for which σy = 2τcr

(5.21)

Equations (5.20) and (5.21) are valid for an applied tensile or compressive stress and can be used in the case of twinning as well. However, τcr for twinning is usually greater than τcr for shear. Some values of the critical resolved shear stress for slip in some common metals, and their temperature dependence, are shown in Figure 5.13. Note that the critical resolved shear stress for HCP and FCC (close-packed) structures rises only modestly at low temperatures, whereas that for the BCC and rock salt structures increases significantly as temperature decreases.

396

MECHANICS OF MATERIALS

Cooperative Learning Exercise 5.3

Consider a single crystal of BCC Mo oriented such that a tensile stress of 52 MPa is applied along a [010] direction. From Table 5.1, we see that the slip plane for BCC Mo is (110) and the slip direction is [111]. Person 1: Determine the angle, φ, between the normal to the slip plane and the applied stress direction. You will find it helpful to draw a (110) plane in a BCC unit cell, and draw the normal to this plane. Once you have determined this direction, you should be able to easily determine the angle it makes with the [010] direction by inspection. Person 2: Determine the angle, λ, between the slip direction and the applied stress direction. You will find it helpful to draw a cubic cell with the [010] and [111] directions as two sides of a triangle, the third side being a face diagonal. At this point, Table 1.8 may be helpful in establishing a relationship between the length of these sides in terms of the lattice parameter, a. You can then use some simple geometry to find λ. Combine your information to calculate the resolved shear stress, τr , using Eq. (5.18). Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, p. 162, 5th ed. Copyright  2000 by John Wiley & Sons, Inc. (a)

x l

s Direction x applied stress [010]

(b)

A

a l

y Slip direction [111]

B a

y

a √2

Normal to slip plane f Slip plane (110)

z

C z

Answers: (see figures) φ = 45◦ ; λ = tan−1 (2)0.5 = 54.7◦ ; τr = 52 MPa(cos 45◦ )(cos 54.7◦ ) = 21.3 MPa

5.1.2.2 Slip in Polycrystals—Strength and Hardening. The yield stress, σy , sometimes called the yield strength, as introduced by Eqs. (5.20) and (5.21), is an important material property, particularly in ductile materials. It not only represents the change from a recoverable, elastic response to a plastic response and the accompanying permanent deformation (see Figure 5.8), but also represents an important design parameter in materials selection. Obviously, when a material undergoes permanent deformation, it may not be useful for its specified purpose and may need to be replaced. As a result, the yield strength may be an even more important mechanical property than the modulus for certain applications. There are no theoretical expressions for the yield strength beyond what has already been described in terms of the critical resolved shear stress in single crystals. This problem is further complicated by the fact that most real materials are polycrystalline—that is, composed of many randomly oriented crystallites, or grains, with corresponding grain boundaries where the grains meet. As a result, only certain grains may be oriented favorably to allow slip to begin, but neighboring grains may not be so oriented, and the stress required to initiate plastic flow increases substantially in polycrystalline materials. It has been shown that at least five independent slip systems

MECHANICS OF METALS AND ALLOYS

103

397

105

W(bcc)

102

Fe(bcc) Co(hcp)

101

Ni(fcc)

103

t0 (psi)

t0 (MN/m2)

104

LiF(NaCl) 100

Cu(fcc) Mg(hcp)

10−1

0

100

200

300

400

102

500

T (K)

Figure 5.13 Temperature variation of critical resolved shear stress for single-crystal metals of different crystal structures. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

must be mutually operative for a polycrystalline solid to exhibit ductility. This means that some of the metals in Table 5.1, particularly the HCP metals such as zinc, do not have sufficient slip systems to undergo significant plastic deformation when in the polycrystalline structure. Since the grain boundaries essentially act as barriers to slip in adjacent crystals (see Figure 5.14), it makes sense that the yield strength should depend on grain size. This is indeed the case, and the Hall–Petch relationship shows an inverse square-root dependence of yield strength on grain size, d: σy = σ0 + kd −1/2

(5.22)

where σ0 and k are empirically determined constants. Equation (5.22) has proved very useful for describing the variation in yield strength with grain size over small ranges of grain sizes, as illustrated for a copper–zinc–brass alloy in Figure 5.15. It is not valid, however, over large size ranges, nor for all grain sizes. The Hall–Petch relationship, and others like it, have found recently renewed interest in the area of nanostructured materials, since they predict enormously large yield strengths as grains approach the nanocrystalline range (d ≈ 10−9 m, or 1 nm). The presence of grain boundaries also affect slip in the material even after plastic deformation has commenced. A phenomenon known as strain hardening, or work hardening or cold working, is the result of constrained dislocation mobility and increased

398

MECHANICS OF MATERIALS

Grain boundary

Slip plane

Grain A

Grain B

Figure 5.14 Illustration of grain boundaries acting as barriers to slip in polycrystalline materials. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, p. 166, 5th ed. Copyright  2000 by John Wiley & Sons, Inc.

Grain size, d (mm) 10−2

5 × 10−3

Yield strength (MPa)

200

30

150 20 100 10

Yield strength (ksi)

10−1

50

0

4

8

12

0 16

d −1/2, (mm−1/2)

Figure 5.15 Influence of grain size on yield strength for a 70–30 Cu–Zn brass alloy. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 167. Copyright  2000 by John Wiley & Sons, Inc.

dislocation density as a material is strained. It is easy to observe this phenomenon; in fact, you have probably used it many times in your life without even knowing it. Simply take a small piece of metal such as a copper tube or an aluminum paper clip, and bend it back and forth several times. In addition to becoming harder to bend and generating heat, the metal will eventually break. Why would a seemingly ductile metal suddenly break? This is strain hardening. As the metal is “worked” back and forth, the dislocations that allow slip and ductility to occur eventually increase in density and decrease in mobility such that the critical resolved shear stress necessary to continue

399

MECHANICS OF METALS AND ALLOYS

plastic flow increases beyond the applied stress. At that point, either you cannot work the piece anymore or brittle fracture (see Section 5.2.2) occurs. The critical resolved shear stress can be related to the dislocation density, ρ (not to be confused with the mass density): τcr = τcr,0 + Aρ 1/2 (5.23) where A is an empirically determined constant that is approximately the product of the shear modulus, G, and the Burger’s vector magnitude, b, and τcr,0 is the critical resolved shear stress in the absence of interfering dislocations. This relationship has been borne out experimentally, as shown in Figure 5.16 for both single- and polycrystalline copper. The dislocation density in a metal increases with deformation or cold work, due to dislocation multiplication or the formation of new dislocations. Consequently, the average distance between dislocations decreases, and the motion of dislocations is limited by the other dislocations around it. Thus, as shown in Eq. (5.23), as the dislocation density increases, the stress necessary to deform the metal increases. In addition to grain size control and strain hardening, a single-phase metal can be strengthened (yield strength increased) by a number of thermal and chemical methods, including solid solution formation and dispersion strengthening. Solute atoms introduced in the formation of solid solutions tend to segregate to dislocations because the total elastic strain energy associated with a dislocation and the solute atom is reduced. As a result, alloys are almost always stronger than pure metals because impurity atoms that go into solid solution ordinarily impose lattice strains (either compressive or tensile, depending on the relative size of the impurity atom) on the surrounding atoms. In this way, lattice strain field interactions between dislocations and these impurity atoms restrict dislocation movement. These same lattice interactions exist during plastic

104

Slope = 1/2

10

1 106

103

102 107

108

109

Dislocation density, r

1010

Resolved shear stress (psi)

Resolved shear stress (MN/m2)

100

1011

(cm/cm3)

Figure 5.16 Resolved shear stress as a function of dislocation density for copper. Data are for  polycrystalline copper; single-crystal copper with one slip system operative; ♦ single-crystal copper with two slip systems operative; and △ single-crystal copper with six slip systems operative. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

Ž

400

MECHANICS OF MATERIALS

deformation, too, so a greater stress is necessary to first initiate and then continue plastic deformation in solid-solution-strengthened alloys. Besides relative size of the solute atoms, the effective elastic modulus of a solute atom also determines its effectiveness as a solid-solution strengthener. The same effect on dislocation movement during slip can be accomplished by two other techniques known as precipitation hardening and dispersion strengthening. In both cases, the principle is the same: Fine dispersions of heterogeneities impede dislocation motion. In precipitation hardening, an alloy is subjected to a solution heat treatment (see Figure 5.17). First, an alloy of suitable composition (line x –x in Figure 5.17) is heated to a temperature above the solvus line but below the solidus line (point 1), and it is held at this temperature to allow complete homogenization of the solid solution. The alloy is quenched to form a supersaturated solution (point 2). The quenched alloy is then reheated and held at this temperature, called the aging temperature (point 3), for a specified period of time to allow the nonequilibrium second phase to precipitate out. Precipitation hardening is a very important method of strengthening many solidsolution alloys such as aluminum and magnesium alloys, copper–beryllium alloys, high nickel-based alloys, and some stainless steels. In dispersion-strengthened alloys, the atmosphere can be controlled during melt processing to form fine compound materials that do not dissolve readily in the base metal. These dispersed particles then act to restrict dislocation movement. A common example is oxide-dispersion-strengthened (ODS) alloys in which oxygen reacts with the parent metal to form higher-melting-point metal oxide particles. These materials are further described in Section 5.4.1. 5.1.2.3 Recovery in Polycrystals—Annealing and Recrystallization. Deformation in polycrystalline materials can lead to not only strain hardening and an increase in dislocation density, as described above, but also a change in grain shape. In some instances, all of these processes can be reversed through heat treatment in what is generically called annealing. Annealing refers to a heat treatment in which a material 800

x Liquid

Temperature (°C)

700 600

Liquid + a 1

a

500

3

400 300 200 100 Al (100%)

1

2

x 2 3

4

5

6

7

8

9

Copper, %

Figure 5.17 Partial Al–Cu equilibrium phase diagram illustrating precipitation hardening. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed., Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

MECHANICS OF METALS AND ALLOYS

401

is exposed to an elevated temperature for an extended period of time and then slowly cooled. The annealing process usually consists of three stages: (1) heating to the desired temperature; (2) holding or “soaking” at that temperature; and (3) cooling at a slow rate, usually to room temperature. Time is an important parameter in each step. During annealing, several processes can occur that lead to what is termed “recovery,” which is a restoration of properties and structures that are more similar to those found in the pre-cold-worked states. We concentrate here on those recovery processes during annealing that are specific to metals and alloys, and we return in later sections to describe annealing in other materials, particularly inorganic glasses. Annealing in metals can first lead to stress relaxation in which stored internal strain energy due to plastic deformation is relieved by thermally activated dislocation motion (see Figure 5.18). Because there is enhanced atomic mobility at elevated temperatures, dislocation density can decrease during the recovery process. At still higher temperatures, a process known as recrystallization is possible, in which a new set of Annealing temperature (°F) 600

600

800

1000

1200

Tensile strength (MPa)

Tensile strength

60

50

500 40

400

30

Ductility (%EL)

400

Ductility 20 300

Recovery

Recrystallization

Grain growth

Grain size (mm)

Cold-worked and recovered grains 0.040

New grains

0.030 0.020 0.010 100

200

300

400

500

600

700

Annealing temperature (°C)

Figure 5.18 Influence of annealing temperature on tensile strength and ductility of a brass alloy. Grain structures during recovery, recrystallization, and grain growth are shown schematically. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 175. Copyright  2000 by John Wiley & Sons, Inc.

402

MECHANICS OF MATERIALS

Table 5.3

Recrystallization and Melting Points of Some Common Metals and Alloys Recrystallization Temperature

Metal Lead Tin Zinc Aluminum (99.999 wt%) Copper (99.999 wt%) Brass (60 Cu–40 Zn) Nickel (99.99 wt%) Iron Tungsten



C

−4 −4 10 80 120 475 370 450 1200



F

25 25 50 176 250 887 700 840 2200

Melting Temperature ◦

C

327 232 420 660 1085 900 1455 1538 3410



F

620 450 788 1220 1985 1652 2651 2800 6170

strain-free and equiaxed (having equal dimensions in all directions) grains are formed. These recrystallized grains are similar to those in the pre-cold-worked metal and have relatively low dislocation densities. The recrystallization process is a diffusion-driven phenomenon similar to the crystallization processes described in Section 3.1.2. As shown in Figure 5.18, the tensile strength decreases and the ductility increases during recrystallization. It is worth reemphasizing that this is a time-dependent phenomenon, too. The degree of recrystallization and recovery depends upon how long the strained material is held at the specified temperature. Typically, the times necessary to obtain complete recrystallization are on the order of hours, and a recrystallization temperature is often defined as the temperature at which recrystallization reaches completion in one hour. The recrystallization temperature is about one-third to one-half of the absolute melting temperature, Tm . For pure metals, the recrystallization temperature is close to 0.3Tm , but for commercial alloys it may be as high as 0.7Tm . Recrystallization temperatures for some common metals and alloys are listed in Table 5.3. 5.1.2.4 Strength and Hardness. One further property that is related to plastic deformation is hardness. Hardness is a measure of a material’s resistance to localized plastic deformation—for example, from a scratch or small indentation. A description of hardness is included in this section because hardness is related to binding forces of molecules, just as strength and modulus are. We also used the term “hardening” to describe the increase in tensile strength as a metal is cold-worked. Hence, hardness is a recoverable quantity during annealing, just as yield strength and ductility are. The principles of strengthening through solid solutions and dispersants also apply here, such that alloys and dispersion-strengthened metals tend to be harder than the pure metals as well. An unfortunate aspect of hardness is that it is difficult to quantify in an absolute manner. Most hardness scales are relative and fairly qualitative, and there is a proliferation of different hardness scales. Some of the more common scales are shown in Figure 5.19. In a typical hardness test, a hard indenter of a standard shape is pressed into the surface of a material under a specified load, causing first elastic and then plastic deformation. The resulting area of the indentation or depth of indentation is measured and assigned a numerical value. The value depends upon the apparatus and scale used.

MECHANICS OF METALS AND ALLOYS

403

10 000 Diamond

10

5000

2000

1000

80

Corundum or sapphire Nitrided steels Topaz

60

File hard

500 40

110 100

20

200 80 100

50

Cutting tools

60

0 Rockwell C

40 20 0 Rockwell B

140 120

Easily machined steels

6

Apatite

5

Fluorite Calcite

4 3

Gypsum

2

Most plastics

40

120 100

5 Brinell hardness

Orthoclase

100

60 130

10

7

Brasses and aluminum alloys

80 20

Quartz

9 8

80 60 40 Rockwell R

20 Rockwell M Talc 1 Mohs hardness

Figure 5.19 Comparison of some common hardness scales. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed., Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

One of the simplest scales is the Mohs scale, in which a material is assigned a place on a scale depending only upon what it will scratch and what will scratch it, relative to the hardest (diamond) and softest (talc) materials known. Since both hardness and strength are relative measures of resistance to plastic deformation, we would expect them to be proportional to each other. This is indeed the

404

MECHANICS OF MATERIALS

Rockwell hardness 60 70 80 90

100 HRB 20

30

40

50 HRC 250

Tensile strength (MPa)

200 Steels 150

1000

100 500

0

Brass

0

Cast iron (nodular)

100

200

300

Tensile strength (103 psi)

1500

50

400

0 500

Brinell hardness number

Figure 5.20 Relationship between hardness and tensile strength for steel, brass and cast iron. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 140. Copyright  2000 by John Wiley & Sons, Inc.

case, but since absolute values of hardness are hard to determine, it is difficult to assign a universal equation to this relationship. The relationship between tensile strength and Brinell hardness is shown in Figure 5.20 for steel, brass, and cast iron. In the past two decades, more standardized measurements of hardness have been made using a technique called nanoindentation. This technique is essentially the same as that described above for traditional hardness testers, except that the indenters are much smaller (about 100 nm), as is the indentation area. Developments in instrumentation have made possible the measurements of loads in the nanonewtons range, and displacements of about 0.1 nm can be accurately measured [1]. The result of a nanoindentation experiment, shown schematically in Figure 5.21b, is a load versus displacement curve, an example of which is shown in Figure 5.21a. In Figure 5.21a, hmax represents the displacement at the peak load, Pmax ; and hc , the contact depth, is defined as the depth of the indenter in contact with the sample under load. The final displacement, hf , is the displacement after complete unloading. The quantity, S, is called the initial unloading contact stiffness and is determined from the slope of the initial portion of the unloading curve, dP /dh, as indicated in Figure 5.21a.

MECHANICS OF METALS AND ALLOYS

405

Pmax

Load, P

Loading

Unloading

S

Possible range for hc

hf hc for ε = 1

hc for e = 0.72

hmax

Displacement, h (a) Surface profile after load removal Indenter h1

h

hs

P

a

Initial surface

Surface profile under load

hc

(b)

Figure 5.21 (a) A typical load-displacement curve and (b) the deformation pattern of an elastic–plastic sample during and after nanoindentation. Reprinted, by permission, from X. Li and B. Bhushan, Materials Characterization, Vol. 48, p. 13. Copyright  2002 by Elsevier Science, Inc.

Nanoindentation hardness, H , is defined as the indentation load divided by the projected contact area of the indentation. From the load–displacement curve, hardness at the peak load can be determined as H =

Pmax A

(5.24)

where A is the projected contact area. The hardness thus has units of N/m2 , or GPa. For an indenter with a known geometry, A is a function of the contact depth, hc . For a perfect triangular-shaped, diamond-tipped indenter known as a Berkovich indenter, the projected contact area is A = 24.56h2c (5.25) Often times corrections must be made to A for imperfect tips and blunting effects.

406

MECHANICS OF MATERIALS

The modulus from a nanoindentation test is often reported in terms of a reduced elastic modulus, Er , to take into account the fact that at this size scale, the elastic response of the probe tip, Ei , as well as the modulus of the test material, E, must be considered: 1 − νi2 1 − ν2 + (5.26) Er = E Ei where ν and νi are Poisson’s ratio for the test material and the indenter material, respectively. For diamond, which is a common indenter material, Ei = 1141 GPa and νi = 0.07. The reduced elastic modulus can be obtained from a load–displacement curve via the stiffness, S, according to S = 2βEr A/π

(5.27)

where A and Er are as before, and β is a constant that depends on the geometry of the indenter. For a Berkovich-type indenter, β = 1.034. There are many experimental details to consider in nanoindentation, but the end result is that both hardness and elastic modulus can be determined relatively routinely on very small samples, including thin films and nanocomposites. What is more, the hardness measurement is quantitative and reproducible. A typical hardness–modulus–indentation depth curve is shown in Figure 5.22 for an Si(100) plane. 5.1.3

Thermomechanical Effects

In addition to the temperature dependence of the properties such as strength and modulus, which we will discuss individually for each material class, there are two fundamental topics that are often described in the context of heat transfer properties or thermodynamics of materials—for example, thermal conductivity or specific heat—but are related more to mechanical properties because they involve dimensional changes. These two properties, thermoelasticity and thermal expansion, are closely related, but will be described separately.

300

E

20

10

0

H

0

100

200

200

100

Elastic modulus (GPa)

Hardness (GPa)

30

0 300

Peak indentation depth (nm)

Figure 5.22 Hardness and elastic modulus as a function of indentation depth at various peak loads for Si(100). Reprinted, by permission, from X. Li and B. Bhushan, Materials Characterization, Vol. 48, p. 14. Copyright  2002 by Elsevier Science, Inc.

MECHANICS OF METALS AND ALLOYS

407

5.1.3.1 Thermal Expansion. Thermal expansion is a measure of the expansion or contraction of a solid when it is heated or cooled. In the most general case, the coefficient of thermal expansion, α, sometimes called the volume thermal expansion coefficient, is defined as the change in volume per unit volume, V , with temperature, T , at constant pressure, P :   1 dV αV = (5.28) V0 dT p

where we use the subscript V to indicate volume thermal expansion. In many materials, the change of dimensions in any one direction may be different from that in the other two directions, depending on the crystallographic directions of the measurement. The change in linear dimensions with temperature along any one direction define the linear thermal expansion coefficient, αL : 1 αL = L0



dL dT



(5.29)

p

For materials that are isotropic, that is, have the same properties in all directions, it can be shown that αV = 3αL . A material that has different properties in different directions is said to be anisotropic. Thus, a linear expansion coefficient, if no direction of measurement is explicitly stated, implies an isotropic material. Conversely, a volume thermal expansion coefficient implies an anisotropic material, and one should exercise caution when deriving linear thermal expansion coefficients from volume-based measurements. Like the melting point, specific heat, and elastic modulus, thermal expansion can be directly related to the binding force between molecules, and thus the potential energy function, as illustrated in Figure 5.23. Just as it affected the energy required to separate bonds during melting or separate bonds under an applied force, the bond strength between atoms affects the ability of atoms to deviate from their equilibrium positions due to increased thermal energy. Due to its origins in lattice vibration, the thermal expansion coefficient shows the same temperature dependence as heat capacity, and at lower temperatures it approaches zero as absolute zero is reached. It is, however, possible for a material to have a negative value of thermal expansion coefficient.

U

r

(a)

U

r

(b)

Figure 5.23 Potential energy diagrams for materials with (a) high melting point, high elastic modulus, and low thermal expansion coefficient and (b) low melting point, low elastic modulus, and high thermal expansion coefficient.

408

MECHANICS OF MATERIALS

5.1.3.2 Thermoelastic Effect. A mechanical phenomenon that involves the thermal expansion coefficient is the thermoelastic effect, in which a material is heated or cooled due to mechanical deformation. The thermoelastic effect is represented by the following relation:   −Vm αET dT = (5.30) dε s CV

where (dT /dε)S is the change in temperature with strain at constant entropy (adiabatic stretching or cooling), Vm is the molar volume, α is the thermal expansion coefficient, E is the tensile modulus, and Cv is the molar heat capacity at constant volume. Most materials have a positive value of α; that is, they expand upon heating, so that (dT /dε)S is negative and the material cools upon straining. When α is negative, the material heats upon straining. This is the case for many polymer elastomers, such as a rubber band, which heats up upon being stretched. 5.1.4

Stress–Strain Diagrams

We turn our attention now to some practical aspects of mechanical property determinations. The important quantities such as modulus, strength, and ductility are typically summarized in graphical form on a stress–strain diagram. The details of how the experiment is performed and how the stress–strain diagram is generated are described for some common types of applied forces below. 5.1.4.1 The Tensile Test. A typical tensile test is performed in what is known as a universal testing apparatus, which is shown schematically in Figure 5.24. The specimen to be tested is typically in the shape of a “dog bone” such that the specimen is wider at the ends than in the center to facilitate mounting in the specimen grips. The sample cross section can be any geometry, as long as it is uniform along the distance between the grips, known as the gauge length. Circular and rectangular cross sections are the most common geometries. The gauge length can be any length, and is not standardized, though gauge lengths of 1 inch (25 mm) and 2 inches (50 mm) are common. One grip, typically the top grip, is attached to a nonmovable load cell that is mounted to the test frame. The load cell measures the force being imparted to the sample. The other grip is attached to a moveable portion of the frame called the crosshead. The crosshead moves downward at a specified rate, called the crosshead speed, which imparts a known extension rate to the sample. These rates are normally very slow, on the order of fractions of an inch per minute. The absolute extension in the sample could be estimated from the extension rate and time, but it is typically measured much more accurately with what is called an extensometer (eks′ ten s¨am′ ter). The tensile test is typically destructive; that is, the sample is extended until it plasticly deforms or breaks, though this need not be the case if only elastic modulus determinations are desired. As described in the previous section, ductile materials past their yield point undergo plastic deformation and, in doing so, exhibit a reduction in the cross-sectional area in a phenomenon known as necking. The “raw data” from a tensile test are the load versus elongation measurements made by the load cell and the extensometer, respectively. To eliminate sample geometry effects, the extension is divided by the initial length to obtain the dimensionless strain (which is occasionally multiplied by 100 and reported as % elongation for samples e

MECHANICS OF METALS AND ALLOYS

409

Load cell

Extensometer Specimen

Moving crosshead

Figure 5.24 Schematic diagram of a tensile test. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

Cooperative Learning Exercise 5.4

Work with a neighbor. The tensile testing machine shown in Figure 5.24 can be used to measure the tensile properties of materials ranging from the most flexible and weak rubber to the strongest and stiffest steel. In order to measure the tensile properties of these vastly different materials, some things on the machine or the way a test is performed may need to be changed. What would you have to change in order to run these two very different types of materials? Name at least three things you can think of. Answers: Load cell; elongation rate; type of grip; type of extensometer. (Courtesy of Brian P. Grady)

that undergo large deformations), and the load is divided by the cross-sectional area to obtain the stress. Stress is plotted versus strain, and the result is a stress–strain diagram. The presence of necking presents a dilemma. The cross-sectional area is integral to the calculation of stress, yet it has a decreasing value as the experiment proceeds, especially where necking is prevalent. Two types of stress are usually encountered in this case. If the actual cross-sectional area corresponding to the load measurement is used, the true stress is obtained. This requires that the cross-sectional area be continuously

410

MECHANICS OF MATERIALS

monitored and matched to the corresponding load measurements. A more convenient method is to divide the measured load by the initial cross sectional area, which is a constant, to obtain what is termed the engineering stress, σeng = F /A0 . There is a corresponding true strain, εtrue , which is the sum of all the instantaneous length changes divided by the instantaneous length (see Figure 5.5), which for small deformations is related to the instantaneous cross-sectional area, A: εtrue



L0 + L = ln L0





A0 ≈ ln A



(5.31)

Obviously, the engineering stress will be less than the true stress (since A0 is always greater than or equal to A), and the engineering strain will be greater than the true strain at a given load for similar reasons. The result is that the true stress–true strain and engineering stress–engineering strain plots look different, as illustrated in Figure 5.25. It is customary to use engineering stress and strain for small deformations, and this is the convention we will use throughout the rest of the book, except where noted.

140

Fracture

130 120

Engineering or true stress (psi) × 103

110 True stress−strain curve

100 90 80 70

Engineering stress−strain curve

60 50

Fracture 40 30 20 10 0

10

20

30

40

50

60

70

80

90

100

Engineering or true strain (in./in. or m/m) × 10−2

Figure 5.25 Comparison of engineering stress-engineering strain and true stress–true strain plots. Reprinted, by permission, from J. F. Shackelford, Introduction to Materials Science for Engineers, 5th ed., p. 192. Copyright  2000 by Prentice Hall, Inc.

MECHANICS OF METALS AND ALLOYS

70

411

U

60

R

Stress, 1000 psi

50

Y

40

L E P

30 20 10 0

0

0.025 0.05 0.075 0.10 0.15 Strain, in./in.

0.2

0.25

Figure 5.26 Stress–strain diagram for mild steel, illustrating different types of stress. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Subscripts will be dropped from stress and strain, and the use of engineering values will be assumed. 5.1.4.2 Modulus and Strength Determinations. Once the stress–strain diagram has been generated, a number of important mechanical properties for a material can be obtained. Consider the stress–strain diagram for mild steel in Figure 5.26. Several points have been labeled for ease of identification. The elastic region of the diagram is easily identifiable. It begins at the origin and is the constant slope region of the graph. Here, the tensile modulus, E, is given by Eq. (5.5), and it can be determined as the slope of the stress–strain curve in the elastic region. Where the elastic region ends is difficult to determine exactly. In Figure 5.26, two points have been labeled that can correspond to the end of the elastic region. The first, point P, is called the proportional limit, and it signals the end of the mathematical relationship between stress and strain that leads to a constant slope—that is, the proportional region. Within the region OP, Hooke’s Law applies. The second point, point E, indicates the elastic limit. Up to point E, the strain is fully recoverable. The implication, however, is that while the response in the region OE may be elastic, there may not be a perfect Hookean response from point P to point E. In reality, it is difficult to differentiate between the proportional and elastic limits for most materials, and either one will produce an adequate determination of the tensile modulus. Beyond point E, the material begins to plasticly deform, and at point Y the yield point is achieved. The stress at the yield point corresponds to the yield strength, σy [see Eq. (5.20)]. Technically, point Y is called the upper yield point, and it corresponds to the stress necessary to free dislocations. The point at which the dislocations actually begin to move is point L, which is called the lower yield point. After point L, the material enters the ductile region, and in polycrystalline materials such as that of Figure 5.26, strain hardening occurs. There is a corresponding increase in the stress

412

MECHANICS OF MATERIALS

required to maintain plastic flow, and a maximum stress value is often achieved in this region, here at point U. The stress at point U is called the ultimate strength, σu , or the ultimate tensile strength (UTS). Sometimes the ultimate tensile strength is a more important design parameter than the yield strength, because it indicates the maximum load that a material will bear, regardless of whether it deforms or not. It should be pointed out that not all materials strain harden, so the ultimate strength can be located at any point along the stress–strain curve; it is simply the maximum strength. We will see examples for which the maximum stress occurs at the yield point, in which case σu = σy . The final point in Figure 5.26 is point R, which represents the breaking point, or rupture point or fracture point, of the material. The stress at point R is called the fracture strength, σf . Again, we will see an example of stress–strain diagrams for which the fracture strength can vary position; it is simply the point at which the material physically separates into two or more pieces. If the material is not extended to the breaking point, then no fracture strength can be determined. 5.1.4.3 Yield Stress and Proof Stress. In Figure 5.26, we noted that there was a distinct point, point P, which indicated the end of the proportional region in the stress–strain diagram, up to which point Hooke’s Law applied. Unfortunately, most ductile metals do not have a true proportional limit, nor a true yield point. A common example of the stress–strain diagram in a tensile experiment is shown in Figure 5.27. Instead of having an upper and lower yield point, the modulus changes slowly and

Yield strength 0.002 offset

Stress, psi

Yield strength 0.0005 offset

O

C

B

Proportional A limit

D

Strain, in./in.

Figure 5.27 Stress–strain diagram indicating method for determining proof stress. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

MECHANICS OF METALS AND ALLOYS

413

continuously until the ductile region is attained. This makes it difficult to determine the proportional limit and modulus, as well as the yield point and yield strength. We must create an artificial convention for determining where the elastic region ends and the ductile region begins. This is achieved with what is known as the proof stress. The proof stress is defined as the stress required to produce a certain definite permanent strain. This is determined by the offset method. A line (DC in Figure 5.27) is drawn parallel to the straight-line portion of the stress–strain diagram (OA in Figure 5.27), starting at a predetermined strain for point D. Typically, this is a value of 0.2% strain (0.002), and the distance OD is known as the “0.2% offset.” The yield point is then taken as the point where the line DC intersects the stress–strain diagram, or point C. The stress at point C is the proof stress, which is then taken as the yield stress and is referred to as the “yield stress at 0.2% offset.” The modulus is still calculated in the straight-line portion of the curve, as indicated by line OA. It should be noted that other types of offsets can be used, such as the 0.05% offset indicated by the line ending at point B. The type of offset used should be clearly indicated in the yield strength determination for these materials. 5.1.4.4 Ductility, Toughness, and Resilience. Several additional useful quantities can be determined from a stress–strain diagram. These are illustrated in Figure 5.28. In addition to the tensile modulus, yield strength, and ultimate strength as described previously, the ductility, which we have previously used in a qualitative sense to describe a metal that undergoes plastic deformation, can be given a quantitative value by reporting the percent elongation at failure. This point is indicated by point 4 in Figure 5.28. Note that in most cases, there is an elastic recovery after fracture (the slope is equal to that given by point 1) that must be accounted for in determining the final elongation. Recall that elongation is   Lf − L0 × 100 (5.32) %Elongation = L0 Stress 3 2

1

5

4

Strain

Figure 5.28 Stress–strain diagram showing (1) modulus, (2) yield strength, (3) ultimate tensile strength, (4) ductility, and (5) toughness. Note the use of proof stress in determination of yield stress. Reprinted, by permission, from J. F. Shackelford, Introduction to Materials Science for Engineers, 5th ed., p. 190. Copyright  2000 by Prentice-Hall, Inc.

414

MECHANICS OF MATERIALS

The % elongation at fracture is thus dependent upon the initial gauge length, which should be cited when reporting values of ductility. Typically, a gauge length of 50 mm (2 in.) is used. Another quantity that is sometimes used to express ductility is the % reduction in area (%RA): %RA =



A0 − Af A0



× 100

(5.33)

Cooperative Learning Exercise 5.5

Work in groups of four. A tensile specimen of 0.505-in. diameter and 2-in. gauge length is subjected to a load of 10,000 lbf that causes it to elasticly deform at constant volume to a gauge length of 2.519 in. Person 1: Calculate the engineering strain. Person 2: Calculate the true strain Person 3: Calculate the % elongation. Person 4: Calculate the percentage reduction in area, %RA. (You may want to consult Person 2 on this.) Compare your answers. How are they similar? How are they different? Persons 1, 3: Determine the engineering stress. Use this value and the engineering strain to estimate the elastic modulus. Person 2, 4: Determine the true stress. Use this value and the true strain to estimate the elastic modulus. Compare your answers. Why are the moduli determined with these two methods the same or different? Answers: ε = (2.519 − 2)/2 = 0.2595; εtrue = ln(2.519/2) = 0.2307; %elongation = (0.2595) × 100 = 25.95%; %RA = (1 − Af /A0 ) × 100 = (1 − 0.79397) × 100 = 20.6. σ = 10,000 lbf /[π(0.505in/2)2 ] = 49.9 kpsi; E = 49.9 kpsi/0.2595 = 192.4 Mpsi σtrue = 10,000(1.259)/[π(0.505in/2)2 ] = 62.9 kpsi; E = 62.9/0.2307 = 272.5 Mpsi. As illustrated in Figure 5.25, the true stress–strain and engineering stress–strain curves can be quite different, especially at high elongations such as this. In fact, relations such as (5.31) are probably not valid at these large elongations. Furthermore, the elastic limit has probably been exceeded at this point, despite our assumptions.

Two remaining quantities, resilience and toughness, are related to the ability of a material to absorb energy. Toughness, as indicated in Figure 5.28, is the area under the stress–strain curve up to the point of fracture. Thus, it is both a measure of a material’s strength and ability to deform plastically. A material with low strength and high ductility may possess more toughness than one with high strength and low ductility. This is an important design consideration in many applications. The continued use of metals in automobile manufacturing (for frames and other structural components) is the result, in part, of safety considerations. There are certainly stronger materials than aluminum, less dense materials than steel, and less costly materials than either, but the combination of properties, including the ability to absorb energy during impact, is what keeps metals attractive for these applications. A quantity related to toughness is the resilience, which is the ability of a material to absorb energy when it is deformed elastically, and then, upon unloading, to recover

MECHANICS OF METALS AND ALLOYS

415

Stress

sy

0.002

∋y

Strain

Figure 5.29 Schematic illustration of how modulus of resilience (shaded area) is determined. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 130. Copyright  2000 by John Wiley & Sons, Inc.

the energy. Resilience is quantified with the modulus of resilience, Ur , which is the strain energy per unit volume required to stress a material from an unloaded state up to the point of yielding. As shown in Figure 5.29, this is simply the area under the curve to the yield point. Again, note how the yield point is determined from the 0.2% offset in metals. Assuming a linear elastic region up to the yield point (which we know is an approximation in most situations), the modulus of resilience can be found from the yield strength, σy , and modulus, E: σy2 (5.34) Ur = 2E 5.1.4.5 Other Types of Stress–Strain Experiments. In addition to the tensile test described above, a number of other mechanical tests can be performed to obtain specific mechanical properties. The compressive modulus, K [see Eq. (5.12)], is obtained from a stress–strain diagram generated in a compression experiment. Although the testing apparatus shown in Figure 5.24 is shown for a tensile test, it can be used for any number of tests, including compression, with the use of proper fixtures. A schematic of a typical compression fixture is shown in Figure 5.30. The sample sits between two platens, which are designed to simply reverse the direction of the applied load, such that the top platen moves downward and the bottom platen moves upward, thus compressing the sample in between them. A stress–strain diagram is again generated, and the compressive modulus, and compressive strength, σc , which is typically the strength at fracture for compression, can be determined. Compression tests are common for load-bearing materials such as cement and concrete, and the devices used to test them can be enormous. Compression tests can, in principle, be conducted with a tensile-type setup in which the sample is mounted between two grips as in Figure 5.24, and the crosshead is moved upward instead of downward. In fact, this is used for some cyclical loading experiments in which tension and compression are alternated at a specific frequency, usually over small deformations. This configuration

416

MECHANICS OF MATERIALS

to load cell

top platen

bottom platen

to crosshead

Figure 5.30 Schematic diagram of compression fixture for testing apparatus shown in Figure 5.24.

can present some practical problems, however, particularly with regard to nonrigid samples and with sample slippage in the grips. Other types of fixtures can be used for other types of tests. Figure 5.31 illustrates a type of test known as a three-point bending test, in which the sample of rectangular or circular cross section is mounted between three knife points as shown. This geometry produces tensile stresses at the bottom of the sample and compressive stresses at the top of the sample. Since most samples subjected to this test (like concrete) are weaker in tension than in compression, the specimen fails—literally breaks into two—following the formation of a nearly vertical crack, called the flexural crack, near the center. The load at failure, Ff , can then be used to calculate the modulus of rupture, also known as the bend strength, fracture strength, or flexural strength, σfs : σfs =

MC I

(5.35)

where M is the maximum moment (Ff L/4), I is the moment of inertia, and c is the distance from the neutral axis to the extreme fiber in tension. The relationships for M, I , and c for both rectangular and circular cross sections, along with the corresponding equations for σfs , are given in Figure 5.31. The flexural strength will depend on sample size, since with increasing size there is an increase in the probability of the existence of a crack-producing flaw, along with a corresponding decrease in the flexural strength. This test, and others like it such as the four-point bend, are common for brittle materials such as ceramics, which we

MECHANICS OF METALS AND ALLOYS

417

Possible cross sections

F

b d

Rectangular Circular

Support

L 2

L 2

R

s = stress = Mc I where M = maximum bending moment

c = distance from center of specimen to outer fibers I = moment of inertia of cross section F = applied load M

c

I

s

Rectangular

FL 4

d 2

bd 3 12

3FL 2bd 2

Circular

FL 4

R

pR 4 4

FL pR 3

Figure 5.31 Schematic illustration of a three-point bend experiment for either rectangular or circular cross sections. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 409, Copyright  2000 by John Wiley & Sons, Inc.

will describe in the next section. A number of other specialized tests exist to measure specific properties such as impact strength and fracture strength. The interested reader is referred to the texts (listed at the end of this chapter) on mechanics of specific materials for further information. 5.1.5

Mechanical Properties of Metals and Alloys

After a somewhat lengthy introduction into the theoretical and experimental aspects of mechanical properties, which we will see apply to all material classes, we come finally to the specifics of mechanical properties of metals and alloys. 5.1.5.1 Temperature Dependence of Strength. The modulus and strength of selected metals and alloys are listed in Appendix 7. However, there are some generalizations that can be made and trends to be observed in the mechanical properties of metals and alloys. At elevated temperatures, the thermal recovery processes described in Section 5.1.2.3 can occur concurrently with deformation, and both strength and strain hardening are consequently reduced. The latter effect results in decreasing the difference between yield and tensile strengths until at sufficiently high temperatures, they are essentially equal. At lower temperatures, temperature has a marked influence on deformation in crystalline materials. Temperature can affect the number of active slip systems in some

418

MECHANICS OF MATERIALS

100 Cu

RA (%)

80

Ta 60 Fe 40

Mo

20

W

0 Ductile-to-brittle transition temperatures

600

W

80

Mo 400

sy (103 psi)

sy (MN/m2)

120

Ta

800

TS for Cu 40

200

Fe Cu

0

0

200

400

600

0

T (K)

Figure 5.32 Variation of percent reduction in area (%RA, top graph) and yield strength (bottom graph) with temperature for selected metals. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

metals, particularly HCP metals (see Section 5.1.2.1), and this determines the ductility of polycrystalline metals such as Zn and Mg. Furthermore, dislocations move less readily as temperature is decreased. As a result, yield and tensile strengths of polycrystalline materials increase and their ductility, measured in percent reduction in area, %RA [see Eq. (5.33)], decreases with decreasing temperature, as illustrated in Figure 5.32. Below a certain temperature known as the brittle-to-ductile transition temperature, metals can behave in a brittle manner; that is, they fracture before they undergo extensive plastic deformation. This is particularly true for BCC metals below

MECHANICS OF METALS AND ALLOYS

419

the brittle-to-ductile transition temperature. However, FCC metals seldom show such a transition and their ductility decreases only slightly with decreasing temperature. This phenomenon is related to the temperature dependence of the critical resolved shear stress previously shown in Figure 5.13. 5.1.5.2 Strain-Rate Dependence of Strength. The plastic behavior of crystalline materials also depends on tensile strain rate, ε˙ . To this point, we have basically assumed that tests are performed at a constant strain rate that does not vary much from experiment to experiment. Strain rate is an important parameter, however, as illustrated in Figure 5.33. Here, the tensile strength (TS, or σu ) for polycrystalline copper is shown both as a function of temperature and strain rate. The temperature dependence is as discussed in the previous section; that is, TS increases with decreasing temperature. The strain rate dependence on tensile strength is more pronounced at elevated temperatures and, as illustrated by the logarithmic coordinates in Figure 5.33, can be described by the relationship σu = K ε˙ m (5.36)

where K is a proportionality constant and m is a temperature-dependent measure of the strain-rate sensitivity of a material. A similar relationship holds for the yield stress, σy . The value of m is usually small at low temperatures but can be appreciable at high temperatures. For FCC metals at low temperatures, m is of the order 0.01. At low temperatures, BCC metals are more strain rate sensitive, but m is still only about 0.1. In the next section, we describe materials that have large strain-rate exponents on the order of 0.5 and that can undergo great extensions.

1000 105

500

5 × 104 200°C C 400°

100

°C 600

50

104

0°C

TS (psi)

TS (MN/m2)

25°C

5 × 103

80

C 0°

0 10

10 10−8

10−6

10−4

10−2

1

102

104

e (sec−1)

Figure 5.33 Dependence of tensile strength on strain rate and temperature for polycrystalline copper. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

420

MECHANICS OF MATERIALS

Cooperative Learning Exercise 5.6 Using the data in Figure 5.33, select two strain rates for which data are provided. Person 1: Determine the value of m in Eq. (5.36) for copper at 1000◦ C. Person 2: Determine the value of m in Eq. (5.36) for copper at 25◦ C. Compare your results. Do they support the claim that strain-rate dependence is more pronounced at elevated temperatures? What would the value of m be if there were no strain-rate dependence?

Answer : In both cases, Eq. (5.36) can be used to obtain ln(σ2 /σ1 ) = m ln(ε2 /ε1 ), from which m(1000◦ ) ≈ 0.1; m(25◦ ) ≈ 0.02. There is indeed more dependence at higher temperatures. The value of m would be zero for no dependence.

5.1.5.3 Superelasticity and Superplasticity. Superplastic materials are polycrystalline solids that have the ability to undergo large uniform strains prior to failure. For deformation in uniaxial tension, elongations to failure in excess of 200% are usually indicative of superplasticity, although several materials can attain extensions greater than 1000%. There are two main types of superplastic behavior: (a) micrograin or microstructural superplasticity and (b) transformation or environmental superplasticity. Micrograin superplasticity is exhibited by materials with a fine grain size, usually less than 10 µm, when they are deformed within the strain rate range 10−5 to 10−1 /s at temperatures greater than 0.5Tm , where Tm is the melting point in degrees Kelvin. Superplastic deformation is characterized by low flow stresses; this, combined with the high uniformity of plastic flow, has led to considerable commercial interest in the superplastic forming of components.

HISTORICAL HIGHLIGHT

Observations of what appeared to be superplastic behavior were initially made in the late 1920s with a maximum of 361% for the Cd–Zn eutectic at 20◦ C and strain rates of ∼10−8 /s and 405% at 120◦ C and a strain rate of ∼10−6 /s. Jenkins [2] also reported a maximum of 410% for the Pb-Sn eutectic at room temperature but at strain rates of ∼10−6 /s. However, the most spectacular of the earlier observations was that by Pearson in 1934 [3]. While working on eutectics, he reported a tensile elongation of 1950% without failure for a Bi-Sn alloy. Following these observations there was little further interest in the Western world in what was clearly regarded as a laboratory curiosity. Later, and unrelated, studies carried out in the USSR discovered the same phenomenon and the term superplasticity

was coined by Bochavar and Sviderskaya in 1945 [4] to describe the extended ductility observed in Zn-Al alloys. The word superplasticity is derived from the Latin prefix super meaning excess and the Greek word plastikos which means to give form to; sverhplastichnost was used in the original Russian language paper of Bochavar and was subsequently translated as superplasticity in Chemical Abstracts (1947). In 1962 Underwood [5] reviewed previous work on superplasticity and it was this paper, along with the subsequent work of Backofen and his colleagues reported in 1966 [6], which was the forerunner of the present expanding scientific and technological interest in superplasticity. Source: http://callisto.my.mtu.edu:591/notes/ intro.html

421

MECHANICS OF METALS AND ALLOYS

The most important mechanical characteristic of a superplastic material is its high strain-rate sensitivity of flow stress, as given previously in Eq. (5.36). For superplastic behavior, m would be greater than or equal to 0.5; for the majority of superplastic materials, m lies in the range 0.4 to 0.8. The presence of a neck in a material subject to tensile straining leads to a locally high strain rate and, for a high value of m, to a sharp increase in the flow stress within the necked region. Hence the neck undergoes strain rate hardening that inhibits its further development. Thus, a high strain-rate sensitivity results in a high resistance to neck development and the high tensile elongations characteristic of superplastic materials. There are two main types of superplastic alloys: pseudo-single phase and microduplex. In the former class of material, a combination of hot and cold working and heat treatment is employed to develop a fine-scale distribution of dispersoids so that on recrystallization the alloy will have a grain size less than 5 µm. Ideally, the dispersion of particles will also prevent any further grain growth during superplastic deformation; however, most of the second phase particles have usually redissolved into solution at the superplastic forming temperatures. The precipitation-strengthened aluminum alloys based on the Al–Cu, Al–Mg, Al–Zn–Mg and the Al–Li and Al–Cu–Li alloys can be included in this group. Some typical properties of the Al-based superplastic alloys are presented in Table 5.4. Other materials include dispersion-strengthened copper alloys

Table 5.4 Properties of Selected Alloys Exhibiting Superplasticity

Alloy Composition

Temperature of Applicability (◦ C)

Superplasticity (% Elongation)

m

400–480 510–530

1800 1400

0.45–0.7 0.5–0.8

510–530

800

0.4–0.6

470–530

1000

0.45

450–550 480–550 500–540

600 670 1000

0.4–0.5 0.4–0.65 0.4–0.6

790–940 950–990

1400 400

0.6–0.8 0.35–0.65

810–930 750–830 940–980 880–970

1600 700 1350 900

0.48–0.65 0.5–0.55 0.4–0.6 0.5–0.7

Al-Based Alloys (Balance Al) 6%Cu–0.4%Zr–0.3%Mg 5.5%Zn–2.0%Mg–1.5% Cu–0.2%Cr 2.7%Cu–2.2%Li–0.7% Mg–0.12%Zr 4.8%Cu–1.3%Li–0.4% Mg–0.4%Ag–0.14%Zr 5%Ca–5%Zn 4.7%Mg–0.7%Mn–0.15%Cr 2.5%Li–1.2%Cu–0.6% Mg–0.1%Zr

Ti-Based Alloys (Balance Ti) 6%Al–4%V 5.8%Al–4%Sn–3.5%Zr–0.5% Mo–0.3%Si–0.05%C 4%Al–4%Mo–2%Sn–0.5%Si 4.5%Al–3%V–2%Fe–2%Mo 14%Al–20%Nb–3%V–2%Mo 6%Al–2%Sn–4%Zr–2%Mo

Source: J. Pilling, Superplasticity on the Web, http://callisto.my.mtu.edu:591/FMPro?-db = sp&-format = sp.html&-view

422

MECHANICS OF MATERIALS

Stress (1000 psi)

100 80 60 40 20 0 0

1

2

3

4

5

6

7

8

Deformation Strain (%)

Figure 5.34 Example of superelasticity in an Ni–Ti alloy. A deformation of 7% is fully recovered. (ShapeMemory Applications, Inc.).

where silicide or aluminide particles are used to limit grain growth. The aluminum alloys, which from a commercial viewpoint are the most important of the pseudosingle-phase materials, can be further subdivided into those which are recrystallized prior to superplastic forming and those which acquire their fine grain structure only after a limited amount of deformation at the forming temperature. The microduplex materials are thermomechanically processed to give a fine grain or phase size. Grain growth is limited by having a microstructure that consists of roughly equal proportions of two or more chemically and structurally different phases. This latter group of materials includes α/β titanium alloys, α/γ stainless steels, α/β and α/β/κ copper alloys, and eutectics. Some properties of selected titanium-based superplastic alloys are presented in Table 5.4. A property related to superplasticity is superelasticity, or psuedoelasticity. Many of the same types of alloys that possess superplasticity also possess superelasticity, particularly the titanium-based alloys, and these are the same alloys that were described in Section 3.1.2.2 as shape memory alloys. Recall that these materials have a transformation temperature (cf. Figure 3.5) below which they are in an austenitic–martensitic state and above which they are in the austenitic state. If the material is tested just above its transformation temperature to austenite, the applied stress transforms the austenite to martensite and the material exhibits increasing strain at constant applied stress; that is, considerable deformation occurs for a relatively small applied stress. When the stress is removed, the martensite reverts to austenite and the material recovers its original shape. This effect makes the alloy appear extremely elastic; in some cases, strains of up to about 10% can be fully recovered (see Figure 5.34). Superelastic shape memory alloys are used in the nearly indestructible frames of eyeglasses, and they makes excellent springs. Many biomedical applications use superelastic wires and tubes, including catheters and guide wires for steering catheters, and orthodontics.

5.2

MECHANICS OF CERAMICS AND GLASSES

In the previous sections, we saw how materials can behave in elastic and ductile manners and how plastic flow past the yield point could be considered a form of destructive failure due to permanent deformation. In this section, we will see that there

MECHANICS OF CERAMICS AND GLASSES

(a)

(b)

423

(c)

Figure 5.35 Schematic illustration of (a) rupture, (b) ductile fracture, and (c) brittle fracture. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 186. Copyright  2000 by John Wiley & Sons, Inc.

are additional types of destructive failure. In particular, we will concentrate on two types of failure known as brittle fracture and creep. These types of failure are common in, but not limited to, brittle materials such as ceramics and glasses. Brittle materials are characterized by their lack of plastic deformation, as illustrated in Figure 5.35. The ductile materials such as metals have a distinct yield point, after which plastic flow begins, and the process of cross-sectional area reduction known as necking occurs (cf. Section 5.1.4.1). The necking phenomenon is illustrated in Figures 5.35a and 5.35b, after which point failure can occur by either highly ductile fracture (known as rupture) or moderately ductile fracture. In brittle materials, little or no plastic deformation precedes the fracture, as in Figure 5.35c. In terms of the stress–strain diagram, brittle materials exhibit primarily an elastic response, with little or no strain beyond the elastic limit prior to fracture (Figure 5.36). The modulus may be higher or lower than a ductile material. As illustrated earlier by point R in Figure 5.26, the stress at fracture is known as the fracture strength, σf . We will devote the next section to a development of theoretical expressions for fracture strength. A description of creep will then be given, followed by an overview of the mechanical properties of glasses and ceramics. 5.2.1

Fracture Mechanics

5.2.1.1 The Molecular Origins of Fracture. Let us first calculate the theoretical fracture strength of a material—that is, the energy required to separate the atomic planes of a solid. Consider pulling a bar of unit cross section, much as we did schematically in Figure 5.10, except that instead of a shear stress, τ , as was used there, we now are applying a tensile stress, σ , in an attempt to pull the atomic planes apart. Cohesive stresses between atomic planes, σ , vary with the change in interplanar separation, x, where x = a − a0 . Here, a0 is the original (equilibrium) interplanar spacing, for which the corresponding stress is zero, and a is the interplanar spacing at any applied stress.

424

MECHANICS OF MATERIALS

Ultimate strength

Material A (ductile)

Stress

Yield point Proportional limit Elastic limit

Material B (nonductile) Yield strength

Proportional limit

Strain

Offset

Figure 5.36 Comparison of typical stress–strain diagrams for ductile (top curve) and brittle (bottom curve) materials. Reprinted, by permission, from S. Somayaji, Civil Engineering Materials, 2nd ed., p. 24. Copyright  2001 by Prentice-Hall, Inc.

The variation of the tensile stress with interplanar separation can be approximated by a sine curve of wavelength λ (see Figure 5.37), much like we did in Figure 5.10 for an applied shear stress [cf. Eq. (5.15)]: 

2πx σ = σc sin λ



(5.37)

The quantity σc is called the theoretical cleavage strength (we will use the terms fracture and cleavage interchangeably here) and it is the maximum stress required to separate, or cleave, the planes. Since cleavage occurs at an interplanar separation of x = λ/2, the work per unit area required to separate planes, also known as the strain energy, is the area under the curve in Figure 5.37 up to that point, Wε =



λ/2

σ dx

(5.38)

0

or, if we substitute the expression for stress from Eq. (5.37) into Eq. (5.38), Wε =



λ/2

σc sin 0



2πx λ



dx =

λσc π

(5.39)

MECHANICS OF CERAMICS AND GLASSES

425

Stress

sc

a0

l/2

x = a – a0 Distance, x

Figure 5.37 Sine-wave representation of stress variation with interatomic separation distance for two atomic planes. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Recall from Section 2.4.1 that there is work associated with creating any new surface. The fracture of the solid at λ/2 creates two new surfaces where none existed before (see Figure 5.35c), so that the work of cohesion between the two surfaces, Wε , is given by Eq. (2.85), Wε = 2γ , where γ is the surface energy of the solid. We can equate this work with the work calculated in Eq. (5.39) to give σc λ = 2γ π

(5.40)

We also know that to a first approximation, Hooke’s Law applies to the elastic region, which for brittle materials is effectively up to the point of cleavage, so that the stress at cleavage, σc , is given by Hooke’s Law as σc = Eεc

(5.41)

where E is the elastic modulus, and εc is the strain at cleavage, (a − a0 )/a0 . The interplanar separation x = a − a0 can be estimated as λ/2 at cleavage, so that Eq. (5.41) becomes Eλ (5.42) σc = 2a0 Solving Eq. (5.40) for λ and substituting into Eq. (5.42) gives σc2 =

Eγ π a0

(5.43)

426

MECHANICS OF MATERIALS

Cooperative Learning Exercise 5.7

Use values of modulus from Appendix 7, along with surface energy from Appendix 4, to estimate the following quantities. In both cases, take an interplanar distance of 10−8 cm as an order-of-magnitude estimate. Person 1: Calculate the σc and the ratio σc /E for MgO. Person 2: Calculate the σc and the ratio σc /E for Au. Compare your answers. Based on these results, what would you expect the theoretical ratio of σc /E to be? Answers: σc (MgO) = [(210 × 109 N/m2 )(0.66 J/m2 )/(10−10 m)]1/2 = 37.3 GPa; σc ≈ E/6 σc (Au) = [(77 × 109 N/m2 )(1.4 J/m2 )/(10−10 m)]1/2 = 32.8 GPa; σc ≈ E/2.5; generally E < σc < E/10

√ or, since π is a numerical factor close to 1 (we are looking for only order of magnitude effects here), we have Eγ σc ≈ (5.44) a0 The functional dependences of Eq. (5.44) are important: The cleavage strength is proportional to the square root of both modulus and surface energy and is inversely proportional to the square root of the equilibrium interplanar distance. These relationships return in subsequent sections. For now, work through the Cooperative Learning Exercise 5.7 to see some important generalizations about the theoretical cleavage strength and modulus. 5.2.1.2 Fracture in Real Brittle Materials—Griffith Theory. In the Cooperative Learning Exercise 5.7, we estimated that the theoretical cleavage strength might lie somewhere between the value for the modulus and E/10. In reality, the fracture strength of most ceramic materials is about E/1000. What is wrong with Eq. (5.44)? We took experimental values for modulus and surface energy, which, granted, may not be very accurate, but are not off by a combined four orders of magnitude necessary to cause these errors. The problem is with the denominator. Just as dislocations caused the theoretical shear strength to differ from the yield strength, there are structural imperfections that cause the experimental fracture strength to differ from the theoretical cleavage strength. It was proposed by Griffith in 1920 that microscopic cracks exist both within and on the surface of all real materials, which are deleterious to the strength of any material that does not possess ductility. The presence of cracks whose longest dimension is perpendicular to the direction of the applied tensile stress gives rise to especially large stress concentrations at the crack tip, where the real localized stress can approach the theoretical strength of the material due to a small area over which it is applied (see Figure 5.38). It can be shown that the maximum stress at the tip of a crack, σm , is

MECHANICS OF CERAMICS AND GLASSES

427

s stress concentration

b a

2a

Figure 5.38 The Griffith model for micro-crack induced brittle failure. Cracks at the surface have a length of a, whereas internal cracks have a length of 2a.

related to the applied tensile stress, σ0 , and the geometry of the crack  1/2 a σm = 2σ0 ρ

(5.45)

where ρ is the radius of curvature of the crack tip, b2 /a, where b is the height of the crack and a is its length as defined in Figure 5.38, and a is the length of a surface crack or half the length of an interior crack (see Figure 5.38). Do not confuse the length of the crack with the interplanar spacing; the same terminology is used for a reason, as will be shown momentarily, but the values are certainly not the same. For a crack to propagate, σm > σc ; that is, the maximum stress at the crack tip must exceed the theoretical cleavage strength. Since the radius of curvature cannot be less than the order of an atomic diameter, σm must always be finite, but can attain σc even though the bulk of the material is under fairly low applied stress. The stress required to cause crack propagation is then the stress that leads to fracture. Using surface energy and work arguments similar to those in the development of the theoretical cleavage strength from last section, it can be shown that the fracture strength according to this crack propagation model of Griffith is σf =



2γ E πa

1/2

(5.46)

where γ is again the surface energy of the material, E is its elastic modulus, and a is now one-half the length of an internal crack, or the length of a surface crack. Notice the similarity in form between Eqs. (5.46) and (5.44). As we surmised, it was not our values of modulus or surface energy that were erroneous in the application of Eq. (5.44), but rather the size parameter. Instead of interplanar separations on the ˚ 10−10 m) as in Eq. (5.44), it is actually cracks, on the order order of angstroms (A, −6 of microns (10 m), that lead to fracture as in Eq. (5.46) and account for the nearly hundred fold (square root of 104 ) difference in the observed and calculated values of

428

MECHANICS OF MATERIALS

the fracture strength estimated from the elastic modulus referred to at the beginning of this section. Equation (5.46) is an important result. It shows clearly that fracture strength decreases as the size of the crack increases, and it also shows that fracture strength can be related directly to surface energy and elastic modulus. The Griffith equation applies only to completely brittle materials. For materials that exhibit some plastic deformation prior to fracture, the surface energy term may be replaced by a term γ + γp , where γp is the plastic deformation energy associated with crack extension. Values of γp can be up to 1000 times greater than γ , such that the latter can often be disregarded for materials that undergo some plastic deformation prior to fracture. Finally, the Griffith equation tells us that if we can reduced crack sizes, or eliminate them altogether, the fracture strength of a material can be manipulated. We will see that this has important applications in glass science.

Cooperative Learning Exercise 5.8

The crack detection limit for a device that inspects steel structural beams is 3 mm. A structural steel with a fracture toughness of 60 MPa · m0.5 has no detectable surface cracks. Assume a geometric factor of unity to determine: Person 1: The load at which this structural steel may undergo fast fracture if a surface crack just at the detection limit goes undetected. Person 2: The % increase in crack size required for this structural steel to undergo fast fracture at an applied load of 500 MPa. Answer : σc = 618 MPa; a = 4.58 mm (52.7% increase).

5.2.1.3 Failure Modes and Fracture Toughness. Once the critical fracture stress has been achieved and the crack begins to propagate, it can do so in one of three primary modes (see Figure 5.39). Mode I is a tensile, or opening, mode. Modes II and III are sliding and tearing modes, respectively. We will limit our present description to the opening mode, though you should be aware that corresponding values for the quantities developed in this section apply to the other two modes as well. The parameter K1 , known as the mode I stress intensity factor, is used to characterize the distribution of stresses at the crack tip and is a function of the applied stress, the crack size (length), and the sample geometry. Stress intensity factors are conveniently defined in terms of the stress components on the crack plane. In most cases, fracture is dominated by tensile loading, so that the important stress intensity factor is K1 . The value of K1 at which fracture occurs is the critical value of the stress intensity factor at which propagation initiates, K1c , also known as the fracture toughness. Do not confuse the fracture toughness with the toughness defined in Section 5.1.4.4, which was the area under the stress–strain curve. The fracture toughness is specific to fracture in brittle materials, and it indicates the conditions of flaw size and stress necessary for brittle fracture. It is defined as √ K1c = Y σc πa (5.47)

where σc is the theoretical cleavage strength, a is crack length as defined in Figure 5.38 (length of a surface crack, or half the length of an internal crack), and Y is a geometric

MECHANICS OF CERAMICS AND GLASSES

I

mode I opening mode (a)

II

mode II sliding mode (b)

429

III

mode III tearing mode (c)

Figure 5.39 The three modes of crack surface displacement: (a) opening mode; (b) sliding mode; and (c) tearing mode.

factor that depends on the sample geometry and the manner of load application. The units of fracture toughness are thus MPa·m1/2 , or psi·in1/2 . The magnitude of K1c diminishes with increasing strain rate, decreasing temperature, and increasing grain size. 5.2.1.4 Fatigue. A failure phenomenon that is closely related to brittle fracture is caused by repeated exposure to cyclic stresses. This type of failure, called fatigue, results in fracture below the ultimate strength of ductile materials, σu , or below the fracture strength, σf , of brittle materials. Even in ductile materials, the result of fatigue is brittle-like failure, insofar as there is little or no plastic deformation prior to failure. This process, like fracture, is caused by the initiation and propagation of cracks. Fatigue is estimated to comprise approximately 90% of all metal failure, though polymers and ceramics are also susceptible to it. Fatigue is particularly important because is can cause catastrophic failure in even the largest of structures due to small, virtually unnoticeable stress fluctuations over long periods of time. Though stresses, especially environmental stresses, are generally not uniform, there are four categories of cyclical stresses to consider. The first is called a reversed stress cycle (Figure 5.40a), in which the stress alternates equally between compressive and tensile loads. A repeated stress cycle occurs when the maximum and minimum stress of a reversed stress cycle are asymmetric relative to the zero stress level, as in Figure 5.40b. Other types of cyclical stresses include a fluctuating stress cycle, which has a maximum and minimum tensile or compressive load (but not both), and a random stress cycle, in which the stresses are neither periodic nor uniform in the magnitudes. In most cases, the stress cycle is characterized by the mean stress, σ = (σmax + σmin )/2, the range of stress, σr = σmax − σmin , and the stress amplitude, S = (σmax − σmin )/2. Note that for a reversed stress cycle, the mean stress is zero, so that the stress amplitude is usually a more meaningful quantity. Fatigue tests are normally performed by applying one of the stress cycles described above until the test specimen fractures. The number of cycles to failure, Nf , at a

MECHANICS OF MATERIALS

Compression Tension −

430

smax

Stress

+

0

smin Time (a)

Compression Tension −

smax S

Stress

+

sr s 0 smin Time (b)

Figure 5.40 Variation of stress with time that accounts for fatigue failure by (a) a reversed stress cycle and (b) a repeated stress cycle. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 210. Copyright  2000 by John Wiley & Sons, Inc.

specified stress amplitude, S, sometimes called the fatigue life, then characterizes the material. For some materials, a fatigue limit is reached (see Figure 5.41), which is defined as the value of S below which no failure will occur, regardless of the number of cycles, N . The fatigue strength, sometimes called the endurance limit, is defined as the stress level at which failure will occur for some specified number value of Nf —for example, 107 cycles. As mentioned earlier, fatigue is the result of crack initiation and propagation. Crack growth under fatigue is estimated by the Paris equation: da = A(K)m dN

(5.48)

where a is once again the crack length as defined earlier (length of a surface crack or half the length of an internal crack), N is the number of cycles (not necessarily to failure), A and m are material-dependent constants, and K is the range of stress intensity factor at the crack tip, Kmax − Kmin . Utilization of Eq. (5.47) shows that √ K = Y (σmax − σmin ) πa

(5.49)

Stress amplitude, S

MECHANICS OF CERAMICS AND GLASSES

431

Fatigue limit

103

104

105

106 107 108 Cycles to failure, N (logarithmic scale)

109

1010

109

1010

Stress amplitude, S

(a)

S1

Fatigue strength at N1 cycles

103

104 Fatigue life at stress S1

107

N1108

Cycles to failure N (logarithmic scale) (b)

Figure 5.41 Stress amplitude versus cycles to failure illustrating (a) fatigue limit and (b) fatigue life. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 212. Copyright  2000 by John Wiley & Sons, Inc.

The values of A and m are determined using the Paris equation by plotting da/dN at various values of K on logarithmic coordinates. The values of da/dN are determined from plots of crack length versus number of cycles, as shown in Figure 5.42. The corresponding crack lengths and stress ranges can then be used to determine K using Eq. (5.49). Typical values of m range between 1 and 6. It should be noted that the Paris equation applies only to a specific region of crack formation. As shown in Figure 5.42, the crack growth rate is initially small, and fatigue cracks can be nonpropagating. At high crack growth rates, crack growth is unstable. Only at intermediate values of crack growth rates does the Paris equation apply.

432

MECHANICS OF MATERIALS

Crack length a

s2 > s1

s1 s2

da dN a1,s2 a1

da dN a1,s1

a0

Cycles N

Figure 5.42 Determination of crack growth rate, da/dN from crack length versus number of cycles data. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 217. Copyright  2000 by John Wiley & Sons, Inc.

5.2.2

Creep

Fatigue is an example of the influence of time on the mechanical properties of a material. Another example of a time-dependent mechanical property is creep. Creep, sometimes called viscoplasticity, is defined as time-dependent deformation under constant stress, usually at elevated temperatures. Elevated temperatures are necessary because creep is typically important only above Tmp /2, where Tmp is the absolute melting point of the material. A typical creep experiment involves measuring the extent of deformation, called the creep strain, ε, over extended periods of time, on the order of thousands of hours, under constant tensile loads and temperature. The resulting plot of creep strain versus time (Figure 5.43) shows the resulting creep rate, ε˙ = dε/dt, which is the slope of the

Creep strain, e

Rupture

Primary

∆e

∆t

Tertiary

Secondary

Instantaneous deformation Time, t

tr

Figure 5.43 Representative creep curve illustrating primary, secondary, and tertiary creep. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 226. Copyright  2000 by John Wiley & Sons, Inc.

MECHANICS OF CERAMICS AND GLASSES

433

curve at a specified time. A typical creep curve exhibits three distinct regions, with different functional dependences of strain on time for each. The first region is called primary creep, or transient creep, in which the primary creep rate, ε˙ p , continuously decreases with time. A power-law relationship can be applied in this region: ε˙ p ∝ t n

(5.50)

where t is time and n is a negative number. This region represents the strain-hardening region of creep. A similar power law expression can be used in the tertiary creep regime to describe the increase in tertiary creep rate, ε˙ t , with time, except that n is now (usually) a positive number. Values of n range from −2 to 20 for tertiary creep, with values of n = 4 being typical. The creep rate increases during tertiary creep until failure, termed rupture, occurs. The stage of creep that usually occupies the longest time is secondary creep, or steady-state creep, which is characterized by a constant creep rate, ε˙ s . In this area, strain hardening is balanced by thermal recovery. The variation of creep with time as a function of both load and temperature is illustrated in Figure 5.44. Arrhenius-type relationships have been developed for steadystate creep as a function of both variables such as 

−Ec ε˙ s = Kσ exp RT m



(5.51)

where σ is the applied stress, R is the gas constant, T is absolute temperature, K and m are material constants, and Ec is called the activation energy for creep.

T3 > T2 > T1 σ3 > σ2 > σ1

Creep strain

T3 or σ3

T2 or σ2

T1 or σ1

T < 0.4Tm

Time

Figure 5.44 Influence of applied stress and temperature on creep behavior. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 227. Copyright  2000 by John Wiley & Sons, Inc.

434

MECHANICS OF MATERIALS

Cooperative Learning Exercise 5.9

A proposed jet engine turbine blade material exhibits a steady-state creep rate of 0.0001 in/in-sec at 500◦ C under a stress of 20,000 psi. The activation energy for creep is 40 kcal/mol. Person 1: At what temperature will this creep rate be doubled at the same stress level? Person 2: At what temperature will this creep rate be tripled at the same stress level? Compare your answers. Agree upon a temperature at which you both estimate these blades will operate, and calculate the corresponding creep rate. Answer : In both instances, use a ratio of Eq. (5.51) at the two different temperatures to solve for T(0.0002) = 521◦ C; T(0.0003) = 534◦ C. In military jet engines, the inlet gas temperatures can be in excess of 1400◦ C. At the stated loads, the creep rate is over 106 times that at 500◦ C!

5.2.3

Mechanical Properties of Ceramics and Glasses

In this section, we describe the elastic, ductile, fracture, fatigue, and creep properties that are specific to certain classes of ceramic and glass materials, with primary emphasis on those mechanical properties that make the material favorable for certain applications from a design point of view. 5.2.3.1 Crystalline Ceramics. The moduli of ceramic materials are generally higher than those for metals, being in the range of 70 to 500 GPa. For example, the elastic modulus as a function of temperature for a common engineering ceramic, polycrystalline Al2 O3 , is shown in comparison to moduli for three common metals in Figure 5.45. Note that the modulus is relatively independent of temperature, which is common in polycrystalline ceramics. The moduli for single crystals can be even higher than for the polycrystalline analogues, depending upon the direction of orientation of the crystal relative to the applied stress. For example, the maximum modulus for MgO lies along the [111] direction and has a value of E[111] = 336 GPa, whereas the minimum value is along the [100] direction, for which E[100] = 245 GPa. As is the case for most polycrystalline materials, the modulus of a polycrystalline sample of MgO lies somewhere in between these two values, around 310 GPa. One may not normally think of ceramics as having significant ductile responses, since they are mostly brittle materials that normally fracture around their yield point, but slip is a definite possibility, especially in single crystal ceramics. As was the case for metals, plastic deformation in crystalline ceramics occurs by a slip dislocation mechanism (cf. Section 5.1.2.1). The slip systems in some common ceramic crystals are shown in Table 5.5. Note that the number of independent slip systems is low compared to the metallic structures of Table 5.3, especially the BCC metals. The yield strength and ductility of single crystals are highly dependent upon temperature and strain rate. These effects are illustrated in Figure 5.46 for single crystal Al2 O3 . Note the presence of both an upper yield stress and lower yield stress, much as there was for metals, and that they both decrease exponentially with increasing temperature and decreasing strain rate. This drop in yield point values can be explained

MECHANICS OF CERAMICS AND GLASSES

435

50

60

40

(1010 N/m2)

30

40

20 MgO 20

Steel

10 Al 0

0

400

800 T (K)

E(106 psi)

Al2O3

W

1200

0 1600

Figure 5.45 Comparison of elastic modulus between polycrystalline Al2 O3 and some common metals as a function of temperature. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc. Table 5.5 Slip Systems in Some Ceramic Crystals

Crystal

Slip System

C (diamond), Si, Ge NaCl, LiF, MgO, NaF NaCl, LiF, MgO, NaF

(111)[110] (110)[110] (110)[110] (001)[110] (111)[110] (111)[110] (001)[110] (110)[001] (001)[110] (001)[110] (110) (111) (0001)[1120] (101)[101] (110)[001] (111)[110] (110)

TiC, UC PbS, PbTe CaF2 , UO2 CaF2 , UO2

C (graphite), Al2 O3 , BeO TiO2 MgAl2 O4

Number of Independent Systems 5 2

Comments At T > 0.5Tm At low temperatures At high temperatures

5 5 3

At high temperatures

3 At high temperatures 5 2 4 5

Source: W. D. Kingery, H. K. Bowen and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc.

MECHANICS OF MATERIALS

20

Fracture X 1260° (2300°) Upper yield stress

Fracture X ~0.1 in./in./min

14 12

16

Stress (103 psi)

10 12 8

Lower yield stress 1270°C (2318°F)

~0.025 in./in./min ~0.01 in./in./min

8

~0.0025 in./in./min 1370° (2498°) 4

0

6

Stress (MN/m 2)

436

4

~0.001 in./in./min

1470° (2678°) 1570° (2858°) 1670° (3038°) 0.01 Elongation (in.)

2

(a)

Figure 5.46 Effect of temperature and strain rate on deformation behavior in single-crystal Al2 O3 . From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

in terms of dislocation multiplication, rather than dislocation movement. Notice again how the modulus is relatively unaffected by either temperature or strain rate. At low temperatures and high strain rates, brittle behavior is observed. Finally, the alloying of single-crystal Al2 O3 with different cations such as Fe, Ni, Cr, Ti, and Mg leads to a strengthening effect, much as precipitation hardening in metals. As was the case for metals, polycrystalline ceramics show a dependence of mechanical properties on grain size, as given earlier by the Hall–Petch relation in Eq. (5.22), which in this case applies to fracture strength, σf : σf = σ0 + kd −1/2

(5.22)

Remember that in brittle materials, fracture occurs as a result of the most severe flaw (crack) in the sample. Hence, the correlations of strength with average grain size must reflect processing techniques that produce samples in which the size of the most severe flaw scales with the average grain size. In cases where the sizes of the initial flaws are limited by the grains and scale with the grain size, the strengths vary with a modified form of Eq. (5.22): σf = kd −1/2 (5.52) This behavior is reflected at larger grain sizes in Figure 5.47, whereas Eq. (5.22) is followed at smaller grain sizes.

MECHANICS OF CERAMICS AND GLASSES

437

s (kpsi) (relative)

15 kpsi

0

0

0 (17) HP, Annealed, Machine, Annealed: s (r,t.) (18) HP, Machine, Annealed; s (r,t.) (18) HP, Machine, Annealed; s (400°C) (18) HP, Machine, Annealed; s (700°C) (18) HP, Machine, Annealed; s (1000°C)

0

0

0

15

30

45

60

75

90

105

d 1/2(cm1/2)

Figure 5.47 Effect of grain size on strength in polycrystalline Al2 O3 for various processing methods (HP = hot pressed, rt = room temperature). From Introduction to Ceramics, by W. D. Kingery, H. K. Bowen and D. R. Uhlmann, Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Typically, the most severe flaws are associated with porosity, surface damage, abnormal grains, or foreign inclusions. The effect of porosity on both modulus and strength is illustrated in Figures 5.48 and 5.49, respectively. The solid curve in Figure 5.48 represents the following equation, which holds for some ceramics: E = E0 (1 − 1.9P − 0.9P 2 )

(5.53)

where E0 is the modulus of the nonporous material, and P is the volume fraction porosity. A similar relationship can be used to estimate the flexural strength of a

MECHANICS OF MATERIALS

60

400

Modulus of elasticity (GPa)

50 300 40 30

200

20 100 10 0 0.0

0.2

0.4

0.6

0.8

Modulus of elasticity (10 6 psi)

438

0 1.0

Volume fraction porosity

Figure 5.48 Effect of porosity on the elastic modulus of Al2 O3 at room temperature. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 413. Copyright  2000 by John Wiley & Sons, Inc.

40

30

200 150

20

100 10 50 0 0.0

0.1

0.2

0.3

0.4

0.5

Flexural strength (10 3 psi)

Flexural strength (MPa)

250

0 0.6

Volume fraction porosity

Figure 5.49 Effect of porosity on the flexural strength of Al2 O3 at room temperature. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 413. Copyright  2000 by John Wiley & Sons, Inc.

porous ceramic, σfs , as a function of porosity: σfs = σ0 exp(−nP )

(5.54)

where σ0 and n are experimentally determined constants. This relationship is represented by the solid line in Figure 5.49. The creep properties of ceramics are of particular importance, especially at high temperatures. In general, the principles of Section 5.2.2 still apply, but the ceramic

MECHANICS OF CERAMICS AND GLASSES

439

Temperature (°C) 10

1200

1400

1600

1800 2000 2200 2400

ZrO2-compression

1

ZrO2-compression

Strain rate at 50 psi (in./in./hr)

10−1

MgO-tension 10−2

Al2O3-compression MgOcompression

10−3

Al2O3-tension

Al2O3-compression 10−4 BeO,MgOcompression 10−5 MgO-tension

10−6 0.70

0.65

0.60

0.55

0.50

0.45

0.40

0.35

1000/T (°K)

Figure 5.50 Low-stress creep rates of several polycrystalline oxides. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

creep mechanisms can be very complex, and factors such as diffusional creep, grain boundary creep, and substructure formation must be considered in a complete description of creep in ceramics. Porosity and grain size are also factors in creep, as they were in modulus and strength determinations. A summary of these effects may be seen graphically in Figure 5.50, in which the creep rates of several polycrystalline oxide ceramics are presented as a function of temperature. All of the values fall within a band, represented by the dashed lines, and variations in creep rate values for a given material (e.g., Al2 O3 in tension or compression) can be attributed to differences in microstructure (pores, grain size, etc.). 5.2.3.2 Glasses. Due to the disordered structure, the modulus and strength of most inorganic glasses are less than those in most crystalline ceramics (see Figure 5.51).

440

MECHANICS OF MATERIALS

40 250

Stress (MPa)

Aluminum oxide 150

20

100

Stress (103 psi)

30

200

10 50 Glass 0

0

0.0004

0.0008

0 0.0012

Strain

Figure 5.51 Typical stress–strain diagrams for aluminum oxide and glass. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 410. Copyright  2000 by John Wiley & Sons, Inc.

However, due to other favorable properties, such as optical transparency and ease of formability, glasses find wide industrial use and, as a result, have been the subject of much investigation. The mechanical properties of silicate- based glasses have been extensively studied due to their widespread use as containers, reinforcing fibers, and optical fibers. Consequently, there is a wealth of information on the influence of many variables on the mechanical properties of these glasses. We shall concentrate on (a) composition and temperature effects on modulus and strength and (b) an important method of strength enhancement in glass fibers. A number of empirical expressions exist for estimating the modulus of multicomponent glasses based upon composition. One particularly useful one that provides an absolute value for the modulus, E, is based upon the density of the glass, ρ, and the packing density, Vt , dissociation energy, Ui , mole %, pi , and molecular weight, Mi , of the components: 0.2ρ Vi p i Ui pi (5.55) E= Mi pi Values for Vi and Ui for a number of metal oxide glass components are listed in Table 5.6. In a similar manner, the compressive modulus, K, and Poisson’s ratio, ν, can be determined using these parameters: K = Vt2



Ui pi

(5.56)

MECHANICS OF CERAMICS AND GLASSES

441

Table 5.6 Factors Used to Determine Elastic Modulus in Multicomponent Glasses Using Eq. (5.55) Oxide

Vi (mole %)

Ui (kJ/cm3 )

Li2 O Na2 O K2 O BeO MgO CaO SrO BaO B2 O3 Al2 O3 SiO2 TiO2 ZrO2 P2 O5 As2 O5 ZnO CdO PbO

8.0 11.2 18.8 7.0 7.6 9.4 10.5 13.1 20.8 21.4 14.0 14.6 15.1 34.8 36.2 7.9 9.2 11.7

80.4 37.3 23.4 125.6 83.7 64.9 48.6 40.6 77.8 134.0 64.5 86.7 97.1 62.8 54.8 41.5 31.8 17.6

Source: H. Scholze, Glass. Copyright  1991 by SpringerVerlag.

and ν = 0.5 −

0.139 Vt

(5.57)

respectively, where ρ Vi p i Vt = Mi pi

(5.58)

Strength in glasses is also composition-dependent, as shown in Figure 5.52. Correlations similar to Eqs. (5.55)–(5.58) for the variation of strength with composition do exist; however, it is the effect of surface cracks that more profoundly influences strength in real glasses. There are two important techniques that can be used to improve the strength in glasses: thermal tempering and chemical tempering. Tempering utilizes the fact that glasses, and crystalline ceramics for that matter, can be strengthened by developing a state of compression on their surfaces. Tempered glasses are useful because failure normally occurs under an applied tensile stress, and failure in ceramics and glasses is almost always initiated at the surface. When a permanent compressive stress, called a residual compressive stress, is placed on a surface, either through thermal or chemical means, the applied stress must first overcome this residual compression before the surface is brought into tension under which failure can occur (see Figure 5.53). Notice that the residual stress is compressive in nature at the surface of the plate and is tensile in the center (shaded areas). When

442

MECHANICS OF MATERIALS

1400

Al2O3

sc

MN/m2 1200 1000

CaO Na2O

800 250 MN/m2 200 sb

ZnO,MgO

B2O3

PbO CaO

Na2O

PbO

B2O3

150

ZnO MgO

Al2O3

100

sts

100 MN/m2 80

CaO

B2O3 Al2O3

PbO

60 40

ZnO, MgO

Na2O 0

10 RmOn

20 Wt% 30

(a)

0

10 RmOn

20

Wt% 30

(b)

Figure 5.52 Change in compressive strength, σc , bending strength, σb , and tensile strength, σt , of a Na2 O–SiO2 glass (20–80%) as SiO2 is replaced proportionately by weight with other oxides. Reprinted, by permission, from H. Scholze, Glass, p. 273. Copyright  1991 by Springer-Verlag.

Residual stress Resultant stress Applied stress

40,000 30,000 20,000 10,000 Compressive

0

10,000 20,000 30,000 40,000 Tensile

Stress (psi)

Figure 5.53 Residual stress, applied stress, and resultant stress distribution for transverse loading of a tempered glass plate. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

MECHANICS OF CERAMICS AND GLASSES

443

a transverse load (dash line) is placed on the plate (recall that a bending-type load creates a tensile force at the bottom and a compressive force at the top), the resultant stress (solid line) creates a large compressive force at the top of the plate, which it can accommodate, but the tensile force at the bottom of the plate is counteracted by the residual compressive stress, such that the resultant stress is only mildly tensile in nature, and well below the fracture stress. A similar situation occurs for glass fibers (cylinders) in tension, which possess compressive residual stresses on their surfaces. In thermal tempering, the residual compressive stresses are introduced into the surface of the glass through thermal treatment. A glass initially at a uniform temperature above the glass transition temperature, Tg , is rapidly cooled, or quenched, to a new temperature by the use of cold air or oil baths. The outside of the glass, which initially cools more rapidly than the interior, becomes rigid while the interior is still molten. The temperature difference between the surface and the midplane of the glass rapidly reaches a maximum upon quenching, and the interior contracts at a greater rate than the rigid outside due to thermal expansion differences. The stresses induced in the later stages of cooling remain. The resulting stress profile is nearly parabolic in shape (see Figure 5.53), with the magnitude of the compressive stresses at the surface being approximately twice the maximum tension in the interior of the glass. Obviously, heat transfer rate and the heat transfer coefficient at the surface play important roles in determining the residual stress distribution. Another way to introduce compressive stresses on the surface of glasses is to change the chemical composition of the surface through the selective replacement of surface atoms in a process called chemical tempering. This process increases the molar volume of the surface relative to the interior. This is normally done through the substitution of large ions for small. Diffusivity is an important consideration in this process, since the ion being inserted into the surface of the glass must have sufficient mobility to displace the existing ion. The most common example of chemical tempering is the replacement of Na+ ions in sodium-silicate glasses with K+ ions. The compressive stress on the surface associated with the fractional change in molar volume, V /V , can be approximated by σ =

E(V /V ) 3(1 − 2ν)

(5.59)

where E is the tensile modulus and ν is Poisson’s ratio. The compression depth of chemically strengthened glass is controllable, but in practice is limited to a few hundred microns. In a related glass processing technique called etching, the surface of the processed glass is treated with an acid such as hydrofluoric acid, HF, which removes a thin, outer layer of the glass. With this procedure, cracks can be eliminated, or at least reduced in length. As the crack length decreases, the strength of the glass increases in accordance with Eq. (5.46). In some instances, residual stresses are not desirable, for example, when the glass is to be used as an optical component. Residual stresses can be removed through a processes called annealing, in which the glass is heated to a uniform temperature near the transformation temperature and held for a specific amount of time. The stresses are resolved due to increased atomic mobility at elevated temperatures. The glass is then slowly cooled through the critical viscosity range at a rate sufficiently slow to keep residual stresses from developing. The glass is then rapidly cooled.

444

MECHANICS OF MATERIALS

5.2.3.3 Cement and Concrete∗ . A great irony of the materials engineering and science disciplines is that the single largest material of construction used in terms of volume—namely, cement—finds little or no technical coverage in an introductory text. To assist in remedying this situation, we present here some introductory information on cements and concretes, with an emphasis on their mechanical properties, so that proper consideration can be given them during materials selection and design. First, we must review some terminology. HISTORICAL HIGHLIGHT

The Romans pioneered the use of hydraulic, or water-cured, cement. Its unique chemical and physical properties produced a material so lasting that it stands today in magnificent structures like the Pantheon. Yet the formula was forgotten in the first few centuries after the fall of the Roman Empire and wasn’t rediscovered until 1824 as Portland cement. One Roman version was based on a burned mixture of two major components: volcanic ash—called pozzolana—from Mount Vesuvius, which destroyed Pompeii and nearby towns in

79; and calcium carbonate, the stuff of seashells, chalk, and limestone. Adding water to these sets off a complex set of chemical reactions that convert the gritty pasty stuff into what is essentially artificial stone. The nineteenth-century rediscovery of Roman cement, the aforementioned Portland cement, is made from a combination of burned limestone, clay, and water. It is the single most heavily used human-made material on earth. A.D.

Source: Stuff, I. Amato, p. 33

The term “cement” refers to a substance that can be used to hold together aggregates such as sand or stone. A cement can, in principle, be any kind of adhesive; it need not be inorganic, and it may be a single compound or a mixture. In this section, we will use the term to refer to a specific inorganic mixture of calcium, aluminum, and silicon oxides known as hydraulic cements. Hydraulic cements set through reaction with water, as opposed to air-set cements that are principally organic and set through a loss of water. Hydraulic cements are so named because they set (initial stiffening) and harden (develop strength) through a series of hydration reactions. The starting materials of limestone or chalk (CaCO3 ), gypsum (CaSO4 ), kaolin (Al2 O3 ), shale or sand (SiO2 ), and slags containing various metal oxides such as MgO and FeO are mixed and burned to form a fused mass called a clinker, which is then ground to form cement powder. During the burning process, decomposition of the carbonates occurs: CaCO3

CaO + CO2

(5.60)

A shorthand notation is often used to designate the oxides (e.g., C for CaO and A for Al2 O3 ; see Table 5.7), and it is also used to designate the compounds formed between the components during heating, such as calcium aluminate (CA) and tricalcium silicate (C3 S): CaO + Al2 O3

CaAl2 O4 ≡ CA

(5.61)

3CaO + SiO2

Ca3 SiO5 ≡ C3 S

(5.62)

MECHANICS OF CERAMICS AND GLASSES

445

Table 5.7 Shorthand Notation of Important Oxides in Cement Compositions C A L S C

CaO Al2 O3 Li2 O SiO2 CO2

F M T H S

Fe2 O3 MgO TiO2 H2 O SO3

N K P f

Na2 O K2 O P2 O5 FeO

The primary components of Portland cement prior to the hydrolysis reactions are C3 S (3CaO–SiO2 , tricalcium silicate); C2 S (2CaO–SiO2 , dicalcium silicate); C3 A (3CaO–Al2 O3 , tricalcium aluminate); and C4 AF (4CaO–Al2 O3 –Fe2 O3 , tetracalcium aluminoferrite). Upon addition of water, the hydration reactions initiate, and the hydraulic cement begins to gain strength. This process is very complex, but the strengthening effect is due basically to the formation of three types of hydration products: colloidal products such as C2 SžxH2 O, which have a size of less than 0.1 µm; submicrocrystalline products such as Ca(OH)2 , Al3+ , Fe3+ , and SO4 2− phases with sizes from 0.1 to 1 µm; and microcrystalline products, primarily of Ca(OH)2 , with particle sizes greater than 1 µm. The most common type of hydraulic cement, Portland cement, usually contains mostly colloidal products. The setting and hardening processes can take many days, and the ultimate strength of the cement may not be realized for years in some instances. This is illustrated in Figure 5.54 for several types of Portland cement, where the compressive strength is plotted as a function of time. Recall that most applications of cement such as buildings and bridges call for compressive loads, such that compressive modulus and strength are the operative mechanical properties.

Compressive strength (psi)

6000

Type II

5000

Type I Type III

Type III

4000 Type I

3000 Type II 2000

7 days

14

28

90

180

1 yr

2 yr

Days/years

Figure 5.54 Compressive strength of concrete as it develops with time. Reprinted, by permission, from S. Somayaji, Civil Engineering Materials, 2nd ed., p. 127. Copyright  2001 by Prentice-Hall, Inc.

446

MECHANICS OF MATERIALS

There are different types of Portland cement, depending upon the intended application. Type I is General Purpose; Type II is Moderate Sulfate-Resistant; Type III is High Early Strength for applications where the cement must set rapidly; Type IV is Low Heat of Hydration, which is required for applications such as dams since the hydration reactions are highly exothermic; and Type V is Sulfate-Resistant. Sulfate resistance is important when using cement that may be exposed to some sulfur-containing groundwaters and seawaters since sulfur tends to expand and disintegrate cement. The material of construction known as concrete is obtained by combining cement with aggregates and admixtures in addition to the water required for hydration. The aggregates take up about 60–75% of the volume of the concrete. Aggregates are classified according to size as either fine (about the size of sand) or coarse (about the size of gravel). The aggregate size and quantity is selected to control the consistency, workability, and ultimate mechanical properties of the concrete. The effect of aggregate size on the compressive strength of concrete is shown in Figure 5.55. The increase in strength that comes from coarser gradation is due solely to a reduction in water requirements for the reduced amount of cement. Other important effects on the ultimate mechanical properties of concrete are cure temperature, air content, water-to-cement ratio, and, of course, the use of reinforcements. 5.2.3.4 Specific Properties. One aspect of ceramics that makes them particularly attractive for structural applications is their low density relative to most structural metals, especially for high-temperature applications. One way to illustrate this is to report mechanical properties such as modulus and strength on a per unit weight basis, resulting in quantities known as specific properties. Hence, the tensile modulus divided by the density is known as the specific modulus, and the ultimate strength divided by the density is known as the specific strength. Table 1.31 shows the specific moduli and strengths of some common ceramics in fiber form, as well as values for some common

Compressive strength (psi)

4000

3000

2000

1000

2

3 Finer

4

5

6

7 Coarser

Fineness modulus of mixed aggregates

Figure 5.55 Effect of aggregate size on compressive strength of concrete. Reprinted, by permission, from S. Somayaji, Civil Engineering Materials, 2nd ed., p. 127. Copyright  2001 by Prentice-Hall, Inc.

MECHANICS OF CERAMICS AND GLASSES

447

1000

Young’s Modulus, E (GPa)

100

10

1.0

0.1

0.01 0.1

1

10 100 Strength sf (MPa)

1000

10,000

Figure 5.56 Design chart of modulus versus strength (see inset for type of strength determination). Reprinted, by permission, from M. F. Ashby, Materials Selection in Mechanical Design, 2nd ed., p. 42. Copyright  1999 by Michael F. Ashby.

metals for comparison. This effect is most clearly seen by utilizing the well-known design charts of Ashby [7]. The use of Ashby’s design charts are elaborated upon in Chapter 8, but we preview their use here with two very important charts. The first is a plot of elastic modulus (Young’s modulus) versus strength (either yield strength, compressive strength, tensile strength, or tear strength, as indicated) in Figure 5.56. Materials generally fall into groups (e.g., engineering ceramics, engineering alloys, etc.) with individual materials listed where possible. Figure 5.57 is the corresponding plot in terms of specific properties; that is, the modulus and strength values of Figure 5.57 have been divided by density. In both charts, materials at the upper right-hand side of the diagram have the best combination of modulus and strength. We select here an extreme example to illustrate the importance of specific properties. Suppose we are looking for a refractory material. Tungsten, W, is a well-known refractory metal with a melting point 3387◦ C. Its region of modulus and strength values are highlighted in Figure 5.56. A similar ceramic material in terms of modulus and strength would be Al2 O3 , also highlighted in Figure 5.55, with a melting point of

448

MECHANICS OF MATERIALS

1000

Specific Modulus E/r (GPa/(Mg/m3))

100

10

1.0

0.1

0.01 0.1

1

10

100

1000

10,000

Specific Strength sf /r (MPa/(Mg/m3))

Figure 5.57 Design chart of specific modulus versus specific strength (see inset for type of strength determination). Reprinted, by permission, from M. F. Ashby, Materials Selection in Mechanical Design, 2nd ed., p. 44. Copyright  1999 by Michael F. Ashby.

around 2050◦ C. Now locate these materials in Figure 5.57. Notice how W has moved downward and to the left relative to Al2 O3 , due solely to its high density (19.3 g/cm3 ) compared to Al2 O3 (3.98 g/cm3 ). Similar arguments apply to the use of polymers, for which densities are even lower. These materials are described in the next section. 5.3

MECHANICS OF POLYMERS

In this section, we describe the mechanical properties of a class of materials that continues to grow in terms of use in structural applications. As issues related to energy consumption and global warming continue to increase demands for lightweight, recyclable materials, the development of new polymers and the characterization of recycled polymers will continue to dominate research and development efforts in this area. In terms of the mechanical behavior that has already been described in Sections 5.1 and Section 5.2, stress–strain diagrams for polymers can exhibit many of the same characteristics as brittle materials (Figure 5.58, curve A) and ductile materials (Figure 5.58, curve B). In general, highly crystalline polymers (curve A) behave in a brittle manner, whereas amorphous polymers can exhibit plastic deformation, as in

MECHANICS OF POLYMERS

449

10 60

A

8

6

40 30

4

B 20

2

10 0

Stress (103 psi)

Stress (MPa)

50

C 0

1

2

3

4 Strain

5

6

7

8

0

Figure 5.58 The stress–strain behavior of brittle polymer (curve A), ductile polymer (curve B), and highly elastic polymer (curve C). Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 475. Copyright  2000 by John Wiley & Sons, Inc.

curve B. We will see that these phenomena are highly temperature-dependent, even more so with polymers than they are with metals or ceramics. We will also see that due to the unique structures of some crosslinked polymers, recoverable deformations at high strains up to the point of rupture can be observed in materials known as elastomers (Figure 5.58, curve C). Despite the similarities in brittle and ductile behavior to ceramics and metals, respectively, the elastic and permanent deformation mechanisms in polymers are quite different, owing to the difference in structure and size scale of the entities undergoing movement. Whereas plastic deformation (or lack thereof) could be described in terms of dislocations and slip planes in metals and ceramics, the polymer chains that must be deformed are of a much larger size scale. Before discussing polymer mechanical properties in this context, however, we must first describe a phenomenon that is somewhat unique to polymers—one that imparts some astounding properties to these materials. That property is viscoelasticity, and it can be described in terms of fundamental processes that we have already introduced. 5.3.1

Viscoelasticity

As the term implies, viscoelasticity is the response of a material to an applied stress that has both a viscous and an elastic component. In addition to a recoverable elastic response to an applied force, polymers can undergo permanent deformation at high strains, just as was the case for metals and some glasses, as described previously. The mechanism of permanent deformation is different in polymers, however, and can resemble liquid-like, or viscous flow, just like we described in Chapter 4. Let us first develop two important theoretical models to describe viscoelasticity, then describe how certain polymers exhibit this important property.

450

MECHANICS OF MATERIALS

30

3

s/MN m−2

=E 20

2

10

1

1

2

3

=h

1

2

e

e

(a)

(b)

3

Figure 5.59 Stress–strain behavior of (a) spring of modulus E and (b) dashpot of viscosity η. Reprinted, by permission, from J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, p. 277, 2nd ed. Copyright  1991 by J. M. G. Cowie.

In both models of viscoelasticity, the elastic response is represented by a Hookean spring of elastic modulus E (Figure 5.59a), and the viscous flow is represented by a piston-like device known as a dashpot of viscosity η (Figure 5.59b). Note that we use the non-Newtonian viscosity, η, rather than the Newtonian viscosity, µ, because we make no assumptions about the viscous behavior of the polymer. We can also generalize these models for both tension and shear by using either the elastic modulus, E, or the shear modulus, G. In fact, we will use these moduli more interchangeably than we have previously. Keep in mind that both the spring and dashpot are completely imaginary devices, much as the “frictionless pulley” of introductory Physics, and are utilized only to assist in the development of the two models of viscoelasticity. 5.3.1.1 The Maxwell Model. The first model of viscoelasticity was proposed by Maxwell in 1867, and it assumes that the viscous and elastic components occur in series, as in Figure 5.60a. We will develop the model for the case of shear, but the results are equally general for the case of tension. The mathematical development of the Maxwell model is fairly straightforward when we consider that the applied shear stress, τ , is the same on both the elastic, τe , and viscous, τv , elements,

τ = τe = τv

(5.63)

since any applied stress is transmitted completely through the spring, once it is extended, to the dashpot below. The total shear strain, γ , is the sum of the elastic strain, γe , and the viscous strain, γv , (5.64) γ = γe + γv since the total displacement is the sum of the displacements in the spring and the dashpot, which act independently. We can take the time derivative of Eq. (5.64) to

MECHANICS OF POLYMERS

451

F

Unloaded Spring element A Shear modulus G Strain

ge = t /G (recovery) (permanent gv = t × t viscous h deformation)

ge = t /G h

Viscous element B (dashpot)

0

Time, t

F (a)

(b)

Figure 5.60 (a) Maxwell spring and dashpot in series model of viscoelasticity and (b) constant stress conditions resulting in time-dependent strain. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

obtain the total strain rate, γ˙ = dγ /dt: γe γv dγ = + dt dt dt

(5.65)

Recall from Eq. (5.10) that the shear strain can be related to the shear stress through the shear modulus, G, according to Hooke’s Law, where we now add subscripts to differentiate the elastic quantities from the viscous quantities: τe = Gγe

(5.10)

This relation can be differentiated with respect to time and solved for the strain rate: dτe /dt dγe = dt G

(5.66)

Recall also from Section 4.0 that the viscous shear rate, γ˙v , can be related to the viscous shear stress through the viscosity, η, according to Newton’s Law of Viscosity, Eq. (4.3): dγv τv = η (4.3) dt where the subscript “v” has once again been added to differentiate the various shear stresses and strains. [In this version of Eq. (4.3), the minus sign has been incorporated into the direction of the shear stress through a change in coordinate system for ease

452

MECHANICS OF MATERIALS

of mathematical manipulation.] Equation (4.3) can be solved for the shear rate and substituted into Eq. (5.65), along with Eq. (5.66), to yield dγ 1 dτe 1 = + τv dt G dt η

(5.67)

The utility of the Maxwell model is that it predicts permanent deformation due to flow at any applied stress. As shown in Figure 5.60b, at zero time a stress is applied and the spring deforms according to Hooke’s Law. Immediately following zero time, the dashpot begins to flow under the restoring forces of the spring, which loses its stored energy to the dashpot for dissipation. If the stress is kept constant, then dτ/dt = 0, and Eq. (5.67) becomes Newton’s Law of Viscosity, and the Maxwell body is indistinguishable from a liquid. Once the stress is removed, the spring recovers, again according to Hooke’s Law, and the permanent deformation that remains is due solely to the viscous flow. Under conditions of constant shear, dγ /dt = 0, Eq. (5.67) becomes an ordinary differential equation, which can be solved by separation of variables and integration using the boundary conditions τ = τ0 at t = 0 and τ = τ at t = t to give the following relation for the shear stress, τ , as a function of time τ = τ0 exp



−Gt η



(5.68)

Under constant shear conditions, Eq. (5.68) predicts that the stress will exponentially disappear, so that after a time t = η/G, the stress will have decayed to 1/e of its original value. This time is referred to as the relaxation time, trel , for the material, so that Eq. (5.68) becomes   −t (5.69) τ = τ0 exp trel This stress relaxation process is important, and we shall revisit it after developing the second model of viscoelasticity. 5.3.1.2 The Kelvin—Voigt Model. A similar development can be followed for the case of a spring and dashpot in parallel, as shown schematically in Figure 5.61a. In this model, referred to as the Kelvin–Voigt model of viscoelasticity, the stresses are additive τ = τe + τv (5.70)

and the strain is distributed γ = γe = γv

(5.71)

Hooke’s Law and Newton’s Law can once again be substituted, this time into Eq. (5.70), to give dγ (5.72) τ = Gγ + η dt

MECHANICS OF POLYMERS

453

F

Spring element A with rigidity G

Unloaded

Strain

Tret

grec gr

Viscous element B with viscosity h

g∞ = t / G

Tret t

0 Time

F (a)

(b)

Figure 5.61 (a) Kelvin–Voigt spring and dashpot in parallel model of viscoelasticity and (b) resulting time-dependent strain. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Under conditions of constant stress, dτ/dt = 0, Eq. (5.72) can be solved to give a relation for the total strain as a function of time:

  τ −Gt γ = 1 − exp (5.73) G η When the Kelvin–Voigt element is stressed under these constant stress conditions, part of the energy is stored in the spring, while the remainder of the energy is dissipated in the dashpot. Eq. (5.73) indicates that when the sample is subjected to the stress, τ , its elastic response is retarded by the viscous element and the strain does not appear instantaneously but increases gradually, asymptotically approaching its final value of γ∞ , the strain at infinite time, as shown in Figure 5.61b. Upon releasing the stress, recovery will occur because of the elasticity of the spring, but it is again retarded by the dashpot. The strain vanishes gradually and the sample slowly returns to its original shape, theoretically at infinite time. This time-dependent relaxation process is an example of anelasticity, which is time-dependent elastic behavior. Although all materials exhibit some anelasticity, it is generally neglected in everything but polymers, where it can be significant. In fact, the entire time-dependent elastic behavior described in this section as viscoelasticity is a form of anelasticity. Practically, the time needed to achieve the equilibrium after removal of the stress is called the retardation time, tret , which is equal to the ratio η/G, as in the Maxwell model. Equation (5.73) can then be written

  −t τ γr = (5.74) 1 − exp G tret where γr is the strain at time t. If the retardation time is small, the recovery will take place at finite time. The Kelvin–Voigt model illustrates the behavior of materials

454

MECHANICS OF MATERIALS

exhibiting a retarded elastic deformation, as occurs in such phenomena as the elastic aftereffect, creep recovery, and elastic memory. When a material is subjected to a tensile or compressive stress, Eqs. (5.63) through (5.74) should be developed with the shear modulus, G, replaced by the elastic modulus, E, the viscosity, η, replaced by a quantity known as Trouton’s coefficient of viscous traction, λ, and shear stress, τ , replaced by the tensile or compressive stress, σ . It can be shown that for incompressible materials, λ = 3η, because the flow under tensile or compressive stress occurs in the direction of stress as well as in the two other directions perpendicular to the axis of stress. Recall from Section 5.1.1.3 that for incompressible solids, E = 3G; therefore the relaxation or retardation times are λ/E. Example Problem 5.2

In Cooperative Learning Exercise 5.9, you studied the effect of temperature on the creep rate of an engine turbine blade. Calculate the viscosity in poise that can be assigned to this material at 500◦ C under a stress of 20,000 psi. Solution For the case of continuous deformation under a constant tensile stress, we can write, by analogy to Newton’s Law of Viscosity [Eq. (4.3)], a relationship between tensile strain rate, ε˙ , tensile stress, σ , and Trouton’s coefficient, λ, as σ = λ˙ε For incompressible materials, Trouton’s coefficient can be related to λ = 3η, so that η=

σ 3˙ε

From CLE 5.9, ε˙ = 0.0001 in./in.-s at 500◦ C, and σ = 20,000 psi, such that η=

(20,000 psi)(6.895×104 dyn/cm2 · psi) 3(0.0001s−1 )

η = 4.6×1012 poise

5.3.1.3 The Four-Element Model∗ . The behavior of viscoelastic materials is complex and can be better represented by a model consisting of four elements, as shown in Figure 5.62. We will not go through the mathematical development as we did for the Maxwell and Kelvin–Voigt models, but it is worthwhile studying this model from a qualitative standpoint. The total deformation in the four-element model consists of an instantaneous elastic deformation, delayed or retarded elastic deformation, and viscous flow. The first two deformations are recoverable upon removal of the load, and the third results in a permanent deformation in the material. Instantaneous elastic deformation is little affected by temperature as compared to retarded elastic deformation and viscous deformation, which are highly temperature-dependent. In Figure 5.62b, the total viscoelastic deformation is given by the curve OABDC. Upon unloading (dashed curve DEFG),

MECHANICS OF POLYMERS

455

F

E1

Maxwell unit

Unloaded εv

Kelvin or Voigt unit

D

Instantaneous recovery = εe

B Strain

h1

εr

E

Delayed recovery εr

A F Permanent deformation εv

εe

E2

h2

C

O

G

t Time

F (a)

(b)

Figure 5.62 (a) Four-element spring and dashpot model of viscoelasticity and (b) resulting strain-dependent time diagram. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

instantaneous elastic, εe , and retarded elastic, εr , strains are recovered, but viscous flow, εv , results in permanent deformation. The four-element model is particularly useful for describing creep in polymers. The initial, almost instantaneous, elongation produced by the application of the stress is inversely proportional to the modulus of the material; that is, a polymer with a low modulus stretches considerably more than a material in the glassy state with a high modulus. This rapid deformation, shown as OA in Figure 5.62b, is followed by a region of creep, A to B, initially fast, but eventually slowing down to a constant rate represented by B to D (or B to C for no unloading). This is the steady-state, or secondary, creep rate of Figure 5.43. As a result of the inverse dependence on modulus and time-dependence of strain, creep data for polymers are usually reported in terms of the creep compliance, defined as the ratio of the relative elongation at a specified time to the stress. Though notations vary, the creep compliance is usually represented by J (t), where for the constant-stress tensile experiment shown in Figure 5.62 we have: J (t) =

ε(t) σ0

(5.75)

Notice that the compliance is inversely proportional to the modulus. Equations for the time-dependent strain can be developed for the four-element model shown above, or for any combination of elements that provide a useful model. The corresponding timedependent compliance can then be determined using Eq. (5.75), or the time-dependent modulus using an analogous equation.

456

MECHANICS OF MATERIALS

40°C 9 60°C 92°C

80°C

100°C

log Er(t ), Pa

8

110°C 112°C Stress relaxation of PMMA

115°C

7

120°C

125°C 6 135°C 0.001

0.01

0.1

1 Time, h

10

100

1000

Figure 5.63 Time-dependent modulus for unfractionated poly (methyl methacrylate). Reprinted, by permission, from F. Rodriguez, Principles of Polymer Systems, 2nd ed., p. 216. Copyright  1982 by Hemisphere Publishing Corporation.

5.3.1.4 Time-Temperature Superposition—The WLF Equation. Recall that the stress relaxation, trel , for the Maxwell model is simply the ratio of the viscosity to the modulus. These values are easily measured in an oscillating-load apparatus, much like that described for fatigue testing in Section 5.2.1.4. The time-dependent tensile modulus for a typical amorphous polymer, poly(methyl methacrylate), measured by stress relaxation at various temperatures is shown in Figure 5.63. We could take a cross section of the modulus-time curve at some fixed relaxation time, say 0.001 hours, 1 hour, and 100 hours, and replot the data as log E versus temperature, as in Figure 5.64. In doing so, we notice two reasonably well-defined plateaus at all relaxation times. The first, at E = 1010 Pa, is the region of glassy polymer behavior. The second plateau, at E = 106 Pa, is the region of rubbery polymer behavior. Between the two plateaus is the glass transition, which occurs at T = Tg . Notice that Tg is dependent on relaxation rate. At still higher temperatures, rubbery flow and then liquid-like flow occur.

MECHANICS OF POLYMERS

457

9

t = 10−3 h

log Er (t ), Pa

t=1h 8

t = 100 h 7

6 40

60

80 100 Temperature (°C)

120

140

Figure 5.64 Modulus–temperature master curve based on cross sections of Figure 5.63. Reprinted, by permission, from F. Rodriguez, Principles of Polymer Systems, 2nd ed., p. 217. Copyright  1982 by Hemisphere Publishing Corporation.

40°

log Er (t ),Pa

9

100°

8

110°

7

115° 135°

6 −10

−5

0

+5

log t, hr at 115°C

Figure 5.65 Modulus–time master curve based on WLF-shift factors using data from Figure 5.63 with a reference temperature of 114◦ C. Reprinted, by permission, from F. Rodriguez, Principles of Polymer Systems, 2nd ed., p. 217. Copyright  1982 by Hemisphere Publishing Corporation.

Another method for obtaining a composite curve is to shift the curves in Figure 5.63 relative to some reference temperature, TR , as shown in Figure 5.65. For example, the log E versus time curve at 115◦ C in Figure 5.63 could be selected as the reference curve; that is, TR = 115◦ C. Curves at 110◦ C, 100◦ C, and 40◦ C could be shifted to the left, for example, and the curve at 135◦ C could be shifted to the right, as illustrated in Figure 5.63. The result is a modulus–time master curve, and the process of shifting the curves relative to a reference point is called the time–temperature superposition principle. This is not simply a graphical exercise. The process, which has been borne out experimentally for a number of amorphous polymers, is carried out by calculating

458

MECHANICS OF MATERIALS

a shift factor, aT , according to the Williams–Lendel–Ferry (WLF) equation: log aT =

−17.44(T − Tg ) 51.6 + T − Tg

(5.76)

The WLF equation holds over the temperature range from Tg to about Tg + 100 K. The constants in Eq. (5.76) are related to the free volume. This is a procedure analogous to the one we used to generate time–temperature–transformation (TTT) diagrams for metallic phase transformations in Section 3.1.2.2.

Cooperative Learning Exercise 5.10

Work in groups of three. The shift factor, aT , in the WLF Equation [Eq. (5.76)], is actually a ratio of stress relaxation times, trel , in the polymer at an elevated temperature, T, relative to some reference temperature, T0 , and can be related via an Arrhenius-type expression to the activation energy for relaxation, Erel as aT = exp

Erel R



1 1 − T T0



(CLE 1.1)

You have developed a new semicrystalline polymer, which has a typical activation energy for relaxation of Erel = 120 kJ/mol. You wish to know the creep compliance for 10 years at 27◦ C. You know that, in principle, you can obtain the same information in a much shorter period of time by conducting your compliance tests at a temperature above 27◦ C. Person 1: Calculate the shift factor and the required test duration if you performed the tests at 32◦ C. Person 2: Calculate the shift factor and the required test duration if you performed the tests at 77◦ C. Person 3: Calculate the shift factor and the required test duration if you performed the tests at 180◦ C. Compare your answers. Which is the most reasonable temperature for running the compliance tests? What are some things that you should check before deciding on a final temperature? Answers: aT (305) = 0.45, 4.5 years; aT (350) = 1.03 × 10−3 , 91 hrs; aT (423) = 0.9 × 10−7 , 27 sec; 77◦ C gives the most reasonable test duration, 32◦ C is too long and 180◦ C may be too short to collect sufficient data. You should determine Tg and/or Tm to make sure that the test is neither exactly on the glass transition temperature, nor above the crystalline melting point.

The time–temperature superposition principle has practical applications. Stress relaxation experiments are practical on a time scale of 101 to 106 seconds (10−2 to 103 hours), but stress relaxation data over much larger time periods, including fractions of a second for impacts and decades for creep, are necessary. Temperature is easily varied in stress relaxation experiments and, when used to shift experimental data over shorter time intervals, can provide a master curve over relatively large time intervals, as shown in Figure 5.65. The master curves for several crystalline and amorphous polymers are shown in Figure 5.66.

MECHANICS OF POLYMERS

459

11 Linear polyethylene

Isotactic polypropylene

Log G (t ), dynes/cm2

10 Branched polyethylene 9

polyisobutylene

Atactic polypropylene

8

7 −15

−10

−5 Log time, hr

0

5

Figure 5.66 Stress–relaxation master curve for several polymers. From F. W. Billmeyer, Textbook of Polymer Science, 3rd ed. Copyright  1984 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

5.3.2

Mechanical Properties of Polymers

To a much greater extent than either metals or ceramics, the mechanical properties of polymers show a marked dependence on a number of parameters, including temperature, strain rate, and morphology. In addition, factors such as molecular weight and temperature relative to the glass transition play important roles that are not present in other types of materials. Needless to say, it is impossible to cover, even briefly, all of these effects. We concentrate here on the most important effects that can affect selection of polymers from a mechanical design point of view. 5.3.2.1 Temperature and Strain Rate Effects. Polymers such as polystyrene and poly(methyl methacrylate), PMMA, with a high tensile modulus at ambient temperatures fall into the category of brittle materials, since they tend to fracture prior to yielding. “Tough” polymers, or those that undergo some deformation prior to failure, can be typified by cellulose acetate. These two categories of behavior are illustrated in Figure 5.67, along with their respective temperature dependences. It can be seen from Figure 5.67 that the effect of temperature on the characteristic shape of the curves is significant. As the temperature increases, both the rigidity and the yield strength decrease, while the elongation generally increases. For cellulose acetate, there is a ductile–brittle transition at around 273 K, above which the polymer is softer and tougher than the hard, brittle polymer below 273 K. For PMMA, the hard, brittle characteristics are retained to a much higher temperature, but it eventually reaches a soft, tough state at about 320 K. Thus, if the requirements of high rigidity and toughness are to be met, the temperature is important.

460

MECHANICS OF MATERIALS

70

248 K

70

273 K

293 K

s/MN m−2

50

50 298 K

30

338 K 353 K 10

20 10 2e (a)

303 K 313 K

323 K

10 0

277 K

30

30 323 K 10 0

333 K 10

20 10 2e (b)

Figure 5.67 Influence of temperature on the stress–strain response of (a) cellulose acetate and (b) poly(methyl methacrylate). Reprinted, by permission, from J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed., p. 283. Copyright  1991 by J. M. G. Cowie.

As was illustrated in Figure 5.46 for a ceramic material, temperature and strain rate have similar, if not opposite, effects on the stress–strain relationship in polymers. Figure 5.68 shows the effect of increasing strain rate on the stress–strain diagram for polyethylene. Compare this effect with decreasing temperature for the polymers in Figure 5.67, where similar effects are observed—that is, increasing yield strength and modulus. Despite the similarities in yield strength behavior with temperature and strain rate between polymers, metals, and ceramics, the mechanisms are quite different, as alluded to earlier. Specifically, the necking of polymers is affected by two physical factors that are not normally significant in metals. The first is the dissipation of mechanical energy as heat, which can raise the temperature in the neck, causing significant softening. The magnitude of this effect increases with strain rate. The second is deformation resistance of the neck, which has a higher strain rate than the surrounding polymer, and results in the strain-rate dependence of yield strength, as shown in Figure 5.68. The relative importance of these two opposing effects depends on the material, the length and thickness of the specimen, and the test conditions, especially the strain rate, which can vary greatly in different forming operations. In semicrystalline polymers such as polyethylene, yielding involves significant disruption of the crystal structure. Slip occurs between the crystal lamellae, which slide by each other, and within the individual lamellae by a process comparable to glide in metallic crystals. The slip within the individual lamellae is the dominant process, and leads to molecular orientation, since the slip direction within the crystal is along the axis of the polymer molecule. As plastic flow continues, the slip direction rotates toward the tensile axis. Ultimately, the slip direction (molecular axis) coincides with the tensile axis, and the polymer is then oriented and resists further flow. The two slip processes continue to occur during plastic flow, but the lamellae and spherullites increasingly lose their identity and a new fibrillar structure is formed (see Figure 5.69). A model has been developed by Eyring to describe the effects of temperature and strain rate on flow stress in polymers. The fundamental idea is that a segment of a polymer must pass over an energy barrier in moving from one position to another in the solid. This is an activated process, similar to the many activated processes that we

MECHANICS OF POLYMERS

Yield stress

461

Increasing rate of elongation or decreasing temperature

Stress

Ultimate stress

Strain

Figure 5.68 Effect of increasing strain rate and decreasing temperature on stress–strain curves for polyethylene. Reprinted, by permission, from F. Rodriguez, Principles of Polymer Systems, p. 249, 2nd ed. Copyright  1982 by Hemisphere Publishing Corporation.

Fibrils

10 nm

Direction of extension 10 nm Fibril boundary crack

Figure 5.69 Fibrillar structure of an oriented polymer crystal as the result of an applied tensile force. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 72. Copyright  1997 by Oxford University Press.

have already described and that was first illustrated in Figure 3.1. There is a forward rate and a reverse rate across the barrier, as for all activated processes. The significant point introduced by Eyring is that the application of a shear stress modifies the barrier height, so that in the direction of stress, the rate of segment jumping, formerly slow, now becomes fast enough to give rise to measurable strain change. The final form of the Eyring equation rests on the reasonable assumptions that the imposed strain rate

462

MECHANICS OF MATERIALS

is proportional to the net rate at which segments jump preferentially in the direction of the shear stress and that the dominant shear stress in a tensile test is the maximum shear stress, which at yield is σy /2. It should be of no surprise that the final Eyring result contains an Arrhenius-type expression based on the activation energy barrier for segment jumping, H : 

−H ε˙ y = ε˙ 0 exp RT



σy V ∗ exp 2RT 



(5.77)

where V ∗ is the activation volume, R is the gas constant, T is absolute temperature, ε˙ y is the strain rate at yield, and ε˙ 0 is a constant. It is usually desirable to investigate the effect of changing strain rate on yield stress, so that Eq. (5.77) is rewritten in the form σy = T



2 V∗



  ε˙ y H + 2.303R log T ε˙ 0

(5.78)

Figure 5.70 shows plots of σy /T against log ε˙ y for polycarbonate for a series of temperatures between 21.5◦ C and 140◦ C, where the Tg of polycarbonate is 160◦ C. Note that the predictions of the Eyring model are obeyed, namely, that σy /T increases linearly with log ε˙ y , and at constant log ε˙ y , σy /T increases with decreasing temperature. Thus, the yield strength is dependent on both strain rate and temperature. The effect of temperature relative to the glass transition, Tg , is illustrated further with the shear modulus, G. The variation of shear modulus with temperature for three common engineering polymers is shown in Figure 5.71. In general, the temperature dependence is characterized by a shallow decline of log G with temperature with 0.25 21.5°C

40°C

sy /T (MPa/K)

0.20

60°C 0.15 80°C 100°C 0.10

120°C 140°C

0.05

−5

−4

−3

Log10 strain rate

−2

−1

0

(s−1)

Figure 5.70 Eyring plot for polycarbonate. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 191. Copyright  1997 by Oxford University Press.

MECHANICS OF POLYMERS

463

1.0

G(GPa: log scale)

Nylon 6 0.1 Natural rubber: 5% Sulfur

PVC

0.01

0.001

−100

0

+100

+200

Temperature (°C)

Figure 5.71 Dependence of shear modulus on temperature for three common engineering polymers: crosslinked natural rubber, amorphous polyvinyl chloride (PVC), and crystalline Nylon 6. The typical use temperatures are indicated by dotted lines. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 154. Copyright  1997 by Oxford University Press.

one or two abrupt drops. The shallow decline is due to thermal expansion effects. As the temperature increases, the molecules in the solid move further apart. This effect occurs in all solids. The abrupt drops in modulus are attributable to viscoelastic relaxation processes due to some type of molecular movement—for example, a side group rotation or backbone movement. Note that the modulus of all three polymers at −200◦ C is in the range of 1 to 5 GPa. For natural rubber, which is slightly crosslinked, the modulus falls abruptly on heating from −50 to −25◦ C. This is the glass-to-rubber transition. The modulus then levels off at a value that depends upon the degree of crosslinking. We will describe the mechanical properties of elastomers in more detail later. For polyvinyl chloride (PVC), a small viscoelastic relaxation occurs at −40◦ C, followed by a major drop at 70◦ C attributable to the glass-to-rubber transition. The Nylon 6 (polycaprolactam) sample has a crystallinity of about 50% and shows a series of relaxations when heated. The major relaxations at 50◦ C and 220◦ C are attributable to the glass-to-rubber and crystalline melting transitions, respectively. This behavior is typical of crystalline polymers, which normally possess several relaxations, whereas amorphous polymers such as PVC typically have only one transition. 5.3.2.2 Molecular Weight and Composition Effects. As we saw in the previous section, degree of crystallinity has an effect on the mechanical behavior and characteristics of polymers. A related factor is molecular weight. To the extent that molecular weight affects the glass transition (see Figure 1.72) and the use temperature relative to Tg , in turn, affects the mechanical properties such as shear modulus (see Figure 5.71), molecular weight has an effect on mechanical properties. The relationship between Tg

464

MECHANICS OF MATERIALS

and the number-average molecular weight can be approximated by Tg = Tg∞ −

k

(5.79)

Mn

where Tg∞ is the glass transition temperature at infinite molecular weight, and k is about 2 × 105 for polystyrene and 3.5 × 105 for atactic poly(α-methyl styrene). The form of this empirical equation is useful for relating a number of polymer properties, including tensile strength, to the number-average molecular weight Property = a −

b

(5.80)

Mn

where a and b are constants, although the actual variation with molecular weight sometimes depends on an average between the number- and weight-average molecular weights. If a polymer exhibits a yield point and then undergoes extensive elongation before tensile failure, its ultimate tensile strength increases with increasing molecular weight. This effect is illustrated in Figure 5.72 for polyethylene, where molecular weight is represented as melt viscosity, since there is a direct correlation between the two (see Figure 4.11). A similar plot can be made for tensile modulus. The effects of crystallinity on tensile strength are also included in Figure 5.72 in terms of the polymer density. As the crystalline content of the polymer rises, so does its density, and there is a trend toward higher tensile strength for more highly crystalline polymers. Taken together, crystallinity and molecular weight influence a number of mechanical properties, including hardness, flexural fatigue resistance, elongation at tensile break, and even impact strength. The chance of brittle failure is reduced by

Ultimate strength, psi 0.95 2000 psi

0.94 Density, g/cm3

1000 psi

1500 psi

2775 2010

0.93

1985

0.92

1100 850

0.91

1200

2175

1945

2355

1720

1975

0.90 0.89

0

1

2

3

4

5

6

7

8

9

Log melt viscosity, poise

Figure 5.72 Dependence of ultimate tensile strength of polyethylene on molecular weight (melt viscosity) and crystallinity (density). From F. W. Billmeyer, Textbook of Polymer Science, 3rd ed. Copyright  1984 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

MECHANICS OF POLYMERS

465

raising molecular weight, which increases brittle strength, and by reducing crystallinity. For amorphous polymers, impact strength is found to depend on the weight-average molecular weight. As the degree of crystallinity decreases with temperature during the approach to the crystalline melting point, Tm , stiffness and yield strength decrease correspondingly. These factors often set limits on the temperature at which a polymer is useful for mechanical purposes. Equation (5.80) implies that the tensile strength of a mixture of polymer components (blend or alloy) can be related to the weight average of the component tensile strengths, σt,i : σt = wi σt,i (5.81)

where wi is the weight fraction of component i. In most real polymer blends, the relationship between composition and mechanical properties is much more complicated than Eq. (5.81) would suggest. As described in Section 2.3.3, the ultimate structure of a blend—that is, one phase or two phase, and the resulting microstructure—depends on how the polymers interact. For example, some mechanical properties of a polycarbonate/poly(methyl methacrylate) blend are shown in Figure 5.73. These blends are invariably two-phase structures. At both 70% and 90% PC, the blends exhibit improved tensile modulus, tensile strength, and yield strength, relative to pure PC, 100

100

Eb st

80

st , sy (MPa)

sy 60

60

40

40

toughness

20

Elongation, Eb (%) Toughness • 10−6 (Nm/m3)

80

20

PC/PMMA (Melt Blends) 0 0

20

40

60

80

0 100

PC Content (%)

Figure 5.73 Toughness, elongation at break, Eb , yield strength, σy , and tensile strength σt as a function of composition for a PC/PMMA blend. Reprinted, by permission, from T. Kyu, J. M. Saldanha, and M. J. Kiesel, in Two-Phase Polymer Systems, L. A. Utracki, ed., p. 271. Copyright  1991 by Carl Hanser Verlag.

466

MECHANICS OF MATERIALS

without sacrificing elongation at break. Hence, the area under the stress–strain curve, or toughness, is greater for these two blends than for pure PC. The toughness enhancement can be attributed to the brittle–ductile transition of the brittle phase, but the diameter of the dispersed phase must be smaller than 1 µm to observe this effect. As the amount of PMMA is further increased (% PC decreases), a significant drop-off in elongation at break and toughness is observed, even though the yield and tensile strengths do not change significantly. This is a reflection of the brittle behavior of PMMA, as first illustrated in Figure 5.67b. This example should serve as an illustration only. The properties of specific polymer–polymer blends are highly variable, and experimentation should give way to generalizations in these cases. As first introduced in Section 4.2.3, polymer foams are formed when the second phase is a gaseous one such as air or CO2 . The primary benefit of structural foams in relation to solid polymers is the increased specific modulus. This benefit can only be realized when the foam has the proper skin-core structure, as illustrated in Figure 5.74. As a result of manipulations during processing, an injected molded polypropylene foam, for example, will have a nonuniform density through its cross section, with larger, nearly spherical gas cells in the center and small, elongated cells nearer the surface. This form of cell morphology reflects the changing shear and thermal history experienced by the foaming polymer during processing. The outer layer, or skin, contains very few pores. Without the proper skin formation, this functionally graded material will not exhibit adequate mechanical response, since it is the outer skin that bears the initial load, particularly in bending tests. This is illustrated in Figure 5.75, where the effect of gas cell size on specific modulus is less of an effect than is the removal of the outer skin. In terms of copolymerization, the addition of a comonomer to a crystalline polymer usually causes a marked loss in crystallinity, unless the second monomer crystallizes isomorphously with the first. Crystallinity typically decreases very rapidly, accompanied by reductions in stiffness, hardness, and softening point, as relatively small amounts (10–20 mol%) of the second monomer are added. In many cases, a rigid, fiber-forming polymer is converted to a highly elastic, rubbery product by such minor

1 MM

Figure 5.74 Skin-core morphology of an injection-molded polypropylene structural foam. Reprinted, by permission, from P. R. Hornsby, in Two-Phase Polymer Systems, L. A. Utracki, ed., p. 102. Copyright  1991 by Carl Hanser Verlag.

MECHANICS OF POLYMERS

467

Specific modulus (MNm/kg)

1.5 Skin and core

1.4

Cellular core

1.3 1.2 1.1 1.0 0.9 0.8

180 340 345 345 345 360 420 440 540 Mean cell diameter (µm)

Figure 5.75 Effect of gas cell size on the specific flexural modulus of 6-mm-thick polypropylene structural foam. Samples were tested at 23◦ C with and without their outer skins. Reprinted, by permission, from P. R. Hornsby, in Two-Phase Polymer Systems, L. A. Utracki, ed., p. 115. Copyright  1991 by Carl Hanser Verlag.

modifications. The dependence of mechanical properties on copolymer composition in systems that do not crystallize results primarily from changes in intermolecular forces as measure by cohesive energy (cf. Section 2.3.2). Higher cohesive energy results in higher stiffness and hardness and generally improves mechanical properties. The effect of temperature on the mechanical properties of amorphous copolymers depends on structure, as shown in Figure 5.76. Polymer A and polymer B are two amorphous homopolymers with mechanical properties as indicated. A random copolymer of the two has a comparable modulus–temperature spectrum to the two homopolymers, exhibiting a nearly three-decade drop in shear modulus, with differences only in the relaxation temperature. The relaxation temperature is high for polymer B, low for polymer A, and intermediate to the two for the copolymer. However, the block and graft versions of the copolymer exhibit two relaxation temperatures, corresponding to the two phases of A and B that form in these copolymers. One relaxation process is typical of the polystyrene phase, while the other is typical of the polybutadiene phase. By varying the proportions of comonomer, it is possible to obtain a low styrene content, rubbery solid with glassy inclusions, or a high styrene content, glassy solid with rubbery inclusions. Finally, the use of low-molecular-weight species to improve flow properties called plasticizers normally reduces stiffness, hardness, and brittleness. Plasticization is usually restricted to amorphous polymers or polymers with a low degree of crystallinity because of the limited compatibility of plasticizers with highly crystalline polymers. Other additives, such as antioxidants, do not affect the mechanical properties significantly by themselves, but can substantially improve property retention over long periods of time, particularly where the polymer is subject to environmental degradation. 5.3.2.3 Elastomers. In addition to the numerous ways we have grouped polymers (cf. Section 1.3.2), polymers can also be grouped according to mechanical behavior into the categories plastics, fibers, and elastomers. Though we will not elaborate upon

468

Figure 5.76 Effect of temperature on shear modulus for random, block, and graft copolymers. Bottom curves are the derivative of the log G curves. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 173. Copyright  1997 by Oxford University Press.

MECHANICS OF POLYMERS

469

Ceramics

Stress

Metals

Conventional plastics

Elastomers

Strain

Figure 5.77 Comparison of idealized stress–strain diagrams for metals, amorphous polymers, and elastomers.

these rather non-specific distinctions here, we simply note that elastomers exhibit some unique mechanical behavior when compared to conventional plastics and metals, as in Figure 5.77. The most notable characteristics are the low modulus and high deformations. Elastomers readily undergo deformation and exhibit large, reversible elongations under small applied stresses. Although a large number of synthetic elastomers are now available, natural rubber must still be regarded as the standard elastomer because of the excellently balanced combination of desirable qualities. The most important synthetic elastomer is styrene–butadiene rubber (SBR), which is used predominantly for tires when reinforced with carbon black. Nitrile rubber (NR) is a random copolymer of acrylonitrile and butadiene and is used when an elastomer is required that is resistant to swelling in organic solvents. Elastomers exhibit this behavior due to their unique, crosslinked structure (cf. Section 1.3.2.2). It has been found that as the temperature of an elastomer increases, so does the elastic modulus. The elastic modulus is simply a measure of the resistance to the uncoiling of randomly oriented chains in an elastomer sample under stress. Application of a stress eventually tends to untangle the chains and align them in the direction of the stress, but an increase in temperature will increase the thermal motion of the chains and make it harder to induce orientation. This leads to a higher elastic modulus. Under a constant force, some chain orientation will take place, but an increase in temperature will stimulate a reversion to a randomly coiled conformation and the elastomer will contract.

470

MECHANICS OF MATERIALS

HISTORICAL HIGHLIGHT

Just as World War I had revealed the U.S. dependence on German organic chemicals and provided both the commercial market and the government support for the mining and chemical industries, World War II played an equally impressive role in promoting the domestic materials industries. The war revealed how critically the nation’s security rested on access to materials. The Japanese advances in the Pacific were particularly damaging. In a short time the United States lost its access to a quarter of its supplies of chromite, three-quarters of its imports of tin, and almost all of its supplies of natural rubber. Following the Japanese blockade of exports from the natural rubber plantations of Malaysia, the United States established the Rubber Reserve Program, a crash program

among the leading rubber and chemical companies, to develop a synthetic rubber substitute. In Germany, I. G. Farben had developed a synthetic rubber called BunaN made with 25% styrene contained in a butadiene chain. The Rubber Reserve Program called upon the nation’s four leading rubber manufacturers and Standard Oil to set up new operations for the production of synthetic rubber based on styrene-butadiene. Because of its early experience in styrene production, Dow Chemical was put in charge of the program and the U.S. production of styrene, which had started at 2 million pounds per month in 1941, was ratcheted up to close to half a billion pounds per year by the end of 1944. Source: Materials Matter, Ball, p. 42

Increasing crosslink density

log(ErTR /T )

9

7 Infinite molecular weight

5

Increasing molecular weight

T

Figure 5.78 Qualitative effects of increasing molecular weight and crosslink density on modulus where TR is any reference temperature. Reprinted, by permission, from F. Rodriguez, Principles of Polymer Systems, 2nd ed., p. 221. Copyright  1982 by Hemisphere Publishing Corporation.

Unlike non-crosslinked polymers—for which mechanical properties vary little with molecular weight alone, as long as the molecular weight exceeds that of a polymer segment—elastomers, and network polymers in general, exhibit an increase in elastic modulus with crosslink density, as shown in Figure 5.78. Recall that a network polymer technically has an infinite molecular weight. For a finite molecular weight, the crosslinks are temporary ones and are due to entanglements.

MECHANICS OF POLYMERS

471

h1

h2

1" 2

Figure 5.79 Izod (notched) impact test. Reprinted, by permission, from F. Rodriguez, Principles of Polymer Systems, 2nd ed., p. 235. Copyright  1982 by Hemisphere Publishing Corporation.

5.3.2.4 Other Polymer Mechanical Properties. In addition to the tension, compression, shear, hardness, fatigue, and creep tests already described, there are a few tests that are common for polymers, although they can be applied to any material, in principle. When polymers are used as films, particularly in packaging applications, their resistance to tearing is important. In one test of tear strength, a specimen is torn apart at a cut made by a sharp blade. Energy is provided by a falling pendulum, and the work done is measured by the residual energy of the pendulum. Tear strength and tensile strength are closely related. An impact test is also performed with a swinging pendulum. The impact strength is measured when the pendulum is allowed to hit the specimen, often at a carefully placed cut called a notch. This type of test is often called a notched-impact test, or an Izod impact test for the specific type of apparatus (see Figure 5.79). In the impact test, the original and final heights of the hammer, h1 and h2 , determine the strength of the sample of thickness d which is held in a vise. For an adhesive bond, the peel strength (normal to the bonded surface) or the shear strength (in the plane of the bonded surface) may be measured. For other applications, the stress supportable in compression before yielding may be the most important parameter, as in a rigid foam, for example. A torsional pendulum (Figure 5.80) is often used to determine dynamic properties. The lower end of the specimen is clamped rigidly and the upper clamp is attached to the inertia arm. By moving the masses of the inertia arm, the rotational momentum of inertia can be adjusted so as to obtain the required frequency of rotational oscillation. The dynamic shear modulus, G′ , can be measured in this manner. A related device is the dynamic mechanical analyzer (DMA), which is commonly used to evaluate the dynamic mechanical properties of polymers at temperatures down to cryogenic temperatures. Abrasion resistance usually takes the form of a scratch test, in which the material is subjected to many scratches, usually from contact with an abrasive wheel or a stream of falling abrasive material. The degree of abrasion can be determined by loss of weight for severe damage, but is usually measured by evidence of surface marring.

472

MECHANICS OF MATERIALS

Counterweight Support wire

Logarithmic amplifier

Electromagnetic angular detection Air bearing

Plotter Rotational moment of inertia Clamp Oven

Log A

Specimen Time Clamp

Figure 5.80 Torsion pendulum for the determination of shear modulus and damping as functions of temperature at frequencies around 1 Hz. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 133. Copyright  1997 by Oxford University Press.

Friction, hardness, and abrasion resistance are related closely to the viscoelastic properties of polymers. 5.4

MECHANICS OF COMPOSITES

The effect of dispersoids on the mechanical properties of metals has already been described in Section 5.1.2.2. In effect, these materials are composites, since the dispersoids are a second phase relative to the primary, metallic matrix. There are, however, many other types of composite materials, as outlined in Section 1.4, including laminates, random-fiber composites, and oriented fiber composites. Since the chemical nature of the matrix and reinforcement phases, as well as the way in which the two are brought together (e.g., random versus oriented), vary tremendously, we shall deal with specific types of composites separately. We will not attempt to deal with all possible matrix–reinforcement combinations, but rather focus on the most common and industrially important composites from a mechanical design point of view.

MECHANICS OF COMPOSITES

473

The starting point for these descriptions will be the law of mixtures, which was first introduced in Eq. (1.62) and which can be used to describe, to a first approximation, the composite property, P , that results from a combination of the reinforcement and matrix properties, Pr and Pm , respectively: Pc = Pr Vr + Pm Vm

(5.82)

where Vr and Vm are the volume fractions of the reinforcement and matrix components, respectively. This relationship will not apply in all instances, but will serve as a starting point for descriptions of specific composite systems. 5.4.1

Mechanical Properties of Particle-Reinforced Composites

5.4.1.1 Dispersion Strengthened Metals and Alloys. In this section, we elaborate upon the dispersion-strengthened metals first described in Section 5.1.2.2. The hard particle constituent is a small component in comparison to the softer matrix phase, seldom exceeding 3% by volume, and the particles are only a few micrometers in size. Dispersion-hardened alloys differ from precipitation-hardened alloys in that the particle is added to the matrix, usually by nonchemical means. As a result, the characteristics of the particles largely control the properties of the composite, particularly with regard to strength. The finer the interparticle spacing for a given alloy series, the better the ductility. Cold working is generally required to achieve high strength levels. The production method for dispersion-strengthened metals has a considerable effect on the properties of the final product. The particular production method depends on the materials involved and includes such techniques as surface oxidation of ultrafine powders, internal oxidation of dilute solid–solution alloys, mechanical mixing of fine metal powders, direct production from liquid metals, and colloidal processing. The surface oxidation of ultrafine powders method is used to form sintered aluminum powder (SAP) alloys, in which each aluminum particle is surrounded by a thin oxide layer. The powder is then compacted, sintered, and cold-worked to form a strong, dispersion-strengthened alloy. The unique characteristics of SAP alloys are their hightemperature strength retention. Instead of softening at a temperature about half the melting point, recrystallization and grain growth are avoided and an appreciable fraction of the material’s low-temperature strength persists up to 80% of its melting temperature (see Figure 5.81). The SAP technique has also been employed in the production of lead, tin and zinc oxide-dispersion strengthened alloys. In the colloidal technique, the size and distribution of a dispersed thoria (ThO2 ) phase is controlled to produce dispersion-strengthened alloys, primarily with nickel as the metallic phase. The so-called TD (thoria-dispersed) nickel has modest strength at room temperature, but retains this strength nearly to its melting point. TD nickel is 3 to 4 times stronger than pure nickel in the 870–1315◦ C range, and oxidation resistance of the alloy is better than that of nickel at 1100◦ C. Other dispersion-strengthened systems include Al2 O3 -reinforced copper, lead, and iron and TiO2 and MgO-reinforced copper. In general, the strength of dispersionstrengthened materials is not particularly impressive at low temperatures relative to the metal phase melting point. The primary benefit, as discussed previously, is in the high-temperature strength retention. This effect is summarized in Figure 5.82 for

474

MECHANICS OF MATERIALS

STRENGTH

WROUGHT ALUMINUM

DISPERSION HARDENED ALUMINUM

100 °F (37.8°C)

600 °F (316 °C) TEMPERATURE

Figure 5.81 Strength of dispersion-strengthened aluminum compared with wrought aluminum. Reprinted, by permission, from M. Schwartz, Composite Materials Handbook, 2nd ed., p. 1.29. Copyright  1992 McGraw-Hill Book Co. 200

100

STRENGTHENING FACTOR (f)

50

Al–Al2O3 S.A.P. Cu–3.5 vol.% Al2O3 (INT. OXID) Cu–1.1 vol.% Al2O3 (INT. OXID) Cu–7.5 vol.% Al2O3 (MECH. MIX) Ni–7.5 vol.% ThO2 (SALT DECOMP) 24S–T3 ALUMINUM ‘INCONEL 700’

20

10

5

2

100 hr RUPTURE STRESS ALLOY 100 hr RUPTURE STRESS PURE METAL T = TEST TEMPERATURE (°K)

f=

TM = MELTING POINT (°K) 1 0.4

0.5

0.6

0.7 T/TM

0.8

0.9

1.0

Figure 5.82 The strength of different dispersion-strengthened alloys relative to the pure metal strength as a function of relative temperature. Preparation techniques include sintered aluminum powder (SAP), internal oxidation, and salt decomposition. Reprinted, by permission, from A. Kelly, Composite Materials, p. 62. Copyright  1966 by American Elsevier, Inc.

MECHANICS OF COMPOSITES

475

a number of dispersion-strengthened materials in comparison to pure aluminum and Inconel metals. 5.4.1.2 Cermets. Recall from Section 1.4.5.2 that a cermet is a particular composite that is similar to a dispersion-strengthened metal, except that the grains are generally larger (cf. Figure 1.78). The metal matrix usually accounts for up to 30% of the total volume, though a wide a variety of compositions are possible. In fact, in oxide-based cermets, either the oxide or the metal may be the matrix phase. For example, the thermal shock resistance of a 28% Al2 O3 –72% Cr is good, but can be greatly enhanced by reversing the proportions of the constituents. In the same system, composition reversal reduces the elastic modulus at room temperature by a factor of 10, from 361 to 32.4 GPa. Other oxide-based cermets include Al2 O3 in stainless steel and in a Cr–Mo alloy. The oxide-based cermets have been extensively applied as a tool material in high-speed cutting of difficult to machine materials. Carbide-based cermets fall into three general categories: tungsten carbide-based, chromium carbide-based, and titanium carbide-based. Tungsten carbide is widely used as a cutting-tool material and is high in rigidity, compressive strength, hardness, and abrasion resistance. Cobalt is added up to 35% by volume to assist in bonding, and the properties of the cermet vary accordingly. As with the oxide-based cermets, increasing binder volume produces such property improvements as improved ductility and toughness. Compressive strength, hardness, and tensile modulus are reduced as binder content increases. Combined with superior abrasion resistance, the higher impact strength cermets result in die life improvements of up to 7000%. Chromium carbide-based cermets offer phenomenal resistance to oxidation, excellent corrosion resistance, relatively high thermal expansion, relatively low density, and the lowest melting point of the stable carbides. Also important is the similarity of thermal expansion to that of steel. Titanium carbide, used principally for high-temperature applications, has good oxidation and thermal-shock resistance, good retention of strength at elevated temperatures, and a high elastic modulus. Because of the relatively poor oxidation resistance of cobalt at elevated temperatures, nickel is more commonly used as the matrix phase. Finally, refractory cermets such as tungsten-thoria and nickel-barium carbonate are used in high-temperature applications. 5.4.1.3 Rubber-Reinforced Polymer Composites. One of the major developments leading to plastics with outstanding toughness has been the production of rubber-modified glassy polymers such as the acrylonitrile butadiene–styrene (ABS) resins made by polymerizing a continuous glassy matrix in the presence of small rubber particles. Only a small amount (5–15%) of rubber is needed to effect this improvement, and the optimum size of the rubber particles is 1–10 µm. The mechanism by which toughness is developed in these materials can be explained in terms of the crack propagation mechanism first described in Section 5.2.1.2. Most glassy thermoplastics undergo a specific type of crack formation and propagation called crazing, in which a network of fibrils connect the two bulk surfaces in a polymeric crack, as illustrated in Figure 5.83. The incorporation of rubber particles into the glassy matrix serves to bridge the microvoids and arrest crazing and crack propagation, resulting in increased toughness. This effect is illustrated in Figure 5.84, in which the ductility is improved in polystyrene (PS) through the addition of small rubber particles to produce high-impact polystyrene (HIPS).

476

MECHANICS OF MATERIALS

Fibrillar bridges Microvoids (a)

Crack (b)

Figure 5.83 Schematic illustrations of (a) a craze showing microvoids and fibrillar bridges and (b) a craze followed by a crack. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 493. Copyright  2000 by John Wiley & Sons, Inc.

PS

40

Stress (MPa)

30 HIPS 20

10

0

0

2

4

6 8 Extension (%)

10

12

Figure 5.84 Stress–strain curves for polystyrene (PS) and high-impact polystyrene (HIPS). Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 200. Copyright  1997 by Oxford University Press.

5.4.2

Mechanical Properties of Fiber-Reinforced Composites

As first described in Section 1.4.2, there are a number of ways of further classifying fiber–matrix composites, such as according to the fiber and matrix type—for example, glass-fiber-reinforced polymer composites (GFRP) or by fiber orientation. In this section, we utilize all of these combinations to describe the mechanical properties of some important fiber-reinforced composites. Again, not all possible combinations are covered, but the principles involved are applicable to most fiber-reinforced composites. We begin with some theoretical aspects of strength and modulus in composites. Recall from Section 1.4, and from Figure 1.75 in particular, that the fiber reinforcement can be classified as either continuous or discontinuous. In the next two sections

MECHANICS OF COMPOSITES

477

we describe the mechanical properties of these two classes of fiber reinforcement in composites, with the emphasis on reinforcement of polymer–matrix composites. Generally, the highest strength and stiffness are obtained with continuous reinforcement. Discontinuous fibers are used only when manufacturing economics dictate the use of a process where the fibers must be in this form—for example, injection molding. We begin our description of discontinuous fiber-reinforced composites with some concepts related to the distribution of stresses that are independent of fiber composition. These developments will be limited to unidirectionally aligned fibers (cf. Figure 1.76)—that is, fibers with their axes aligned. First, we will concentrate on continuous, unidirectional fibers, and then the strength and modulus of discontinuous, unidirectional fiber-reinforced composites will be analyzed. 5.4.2.1 Fiber–Matrix Coupling. Consider the case of fibers that are so long that the effects of their ends can be ignored. Furthermore, consider the case where these continuous fibers are all aligned along a preferred direction, as illustrated in Figure 5.85. Three orthogonal axes, 1, 2, and 3, are defined within the composite. When loads are applied to the composite block, we can assume that the block deforms as if the stress and strain are uniform within each component. For example, if a stress σ1 is applied parallel to axis 1, as in Figure 5.86a, the fibers and matrix are approximately coupled together in parallel; that is, the fibers and matrix elongate equally in the direction of σ1 . The fiber and matrix axial strains along axis 1—εf 1 and εm1 , respectively—equal the total strain in the composite:

εf 1 = εm1 = ε1

(5.83)

Fibers extend the entire length of the composite, so that at any section the area fractions occupied by fibers and matrix equal their respective volume fractions, Vf and Vm = 1 − Vf . The total stress, σ1 , must then equal the weighted sum of stresses in fibers and matrix, σf 1 , and σm1 , respectively: σ1 = Vf σf 1 + (1 − Vf )σm1

(5.84)

3

1 2

Figure 5.85 Definition of axes in a continuous, unidirectional fiber-reinforced composite. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 258. Copyright  1997 by Oxford University Press.

478

MECHANICS OF MATERIALS

s2 Fibers

s1

s1

Matrix s2 (a)

(b)

Figure 5.86 Fiber–matrix coupling in continuous, unidirectional fiber-reinforced composites: (a) Coupling for tensile stress parallel to fiber axis. (b) Coupling for stress perpendicular to fiber axis. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 258. Copyright  1997 by Oxford University Press.

This is the rule of mixtures for stress, as given in Eq. (5.82). Fibers and matrix are assumed to carry pure axial tension, with no stress in the 2–3 plane; that is, σf 2 = σf 3 = σm2 = σm3 = 0. We treat the fibers and matrix as acting in a purely elastic manner and neglect the viscoelastic effects of the polymer matrix. Hooke’s Law then applies to both the fiber and the matrix σ f 1 = E f εf 1

(5.85)

σm1 = Em εm1

(5.86)

as well as to the composite, where E1 is the composite modulus σ 1 = E 1 ε1

(5.87)

Substitution of the stresses from Eqs. (5.86) to (5.87) into Eq. (5.84) and division by ε1 yields a prediction for the axial composite tensile modulus: E1 = Vf Ef + (1 − Vf )Em

(5.88)

This, too, is the simple rule of mixtures of Eq. (5.82) applied to modulus. Care must be exercised to use the correct fiber modulus, E1 in the case of anisotropic fibers such as carbon and Kevlar . In Eq. (5.88), it is the axial tensile modulus that must be used. The second term in Eq. (5.88) usually makes only a small contribution, since Em ≪ Ef . Except at very low fiber volume fractions, E1 is then given to a good approximation by E 1 ≈ Vf E f

(5.89)

In contrast, when a stress acts in the 2–3 plane, as in Figure 5.86b, the matrix plays a crucial load-bearing role. Fibers and matrix now couple approximately in series, and the whole tensile force is assumed to be carried fully by both the fibers and matrix. The tensile forces in the fiber and matrix, σf 2 and σm2 , are therefore equal to each other and to the overall stress in the composite, σ2 : σf 2 = σm2 = σ2

(5.90)

MECHANICS OF COMPOSITES

479

Displacement in the fibers and matrix in the direction of σ2 can be approximated as being additive. The total strain is therefore the weighted sum of strains in the fibers and matrix: (5.91) ε2 = Vf εf 2 + (1 − Vf )εm2 Again, the stress in each component is approximated as being pure uniaxial tension; that is, there is no stress in the 1–3 plane, and σf 1 = σf 3 = σm1 = σm3 = 0. Once again, Hooke’s Law applies to the fiber, matrix, and composite, and when the strains from these relationships are substituted into Eq. (5.91), an expression for the transverse tensile modulus of the composite, E2 is obtained: E2 =

Ef Em (1 − Vf )Ef + Vf Em

(5.92)

The axial and transverse tensile moduli for a continuous, unidirectional glass-fiberreinforced epoxy matrix composite as predicted by Eqs. (5.88) and (5.92) are given as a function of volume fraction fiber, Vf , in Figure 5.87. Since Ef ≫ Em , Eq. (5.92) reduces to the approximate expression: E2 ≈

Em 1 − Vf

(5.93)

which indicates that fibers now play the subordinate role for transverse tensile loads in continuous, unidirectional fiber-reinforced composites. Similar arguments can be applied to Poisson’s ratio, and the free contraction parallel to axis 2 when a tensile stress is applied parallel to axis 1, ν12 , is found to obey the rule of mixtures (5.94) ν12 = Vf νf + (1 − Vf )νm The other Poisson’s ratio, ν21 applying to the 1–2 plane does not equal ν12 , but can be found from ν12 , E1 , and E2 . When a shear stress acts parallel to the fibers, τ12 , the composite deforms as if the fibers and matrix were coupled in series. Hence, the shear strain, γ12 , can be found from a relation similar to Eq. (5.91), and the corresponding shear modulus for the composite, G12 , is Gf Gm (5.95) G12 = (1 − Vf )Gf + Vf Gm The same models of mechanical coupling can be used to predict the coefficient of linear thermal expansion in the composite, α1 , based on the moduli, and thermal expansion coefficients of the fiber and matrix, αf and αm , respectively: α1 =

Vf Ef αf + (1 − Vf )Em αm Vf Ef + (1 − Vf )Em

(5.96)

Here, αf is assumed to be independent of direction within the fiber, but for highly anisotropic fibers such as carbon and Kevlar , allowances must be made for their variation in thermal expansion coefficient with direction. The thermal expansion coefficient perpendicular to the fiber axis, α2 , can also be derived: α2 = Vf αf (1 + νf ) + (1 − Vf )αm (1 + νm ) − α1 ν12

(5.97)

480

MECHANICS OF MATERIALS

70

60

Tensile modulus (GPa)

50

40

30

E1

20

E2

10

0

0

0.2

0.4

0.6

0.8

1.0

Volume fraction of fibers

Figure 5.87 Predicted tensile moduli for continuous, unidirectional glass-fiber-reinforced epoxy matrix composite. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 259. Copyright  1997 by Oxford University Press.

It is usually the case that αf ≪ αm and Ef ≫ Em , such that Eqs. (5.96) and (5.97) predict widely differing coefficients of thermal expansion parallel and perpendicular to the fibers. This effect is illustrated in Figure 5.88, where the thermal expansion coefficients in the axial and transverse direction for a continuous, unidirectional glassfiber-reinforced epoxy matrix composite are plotted as a function of fiber volume fraction. It is clear that the fibers are highly effective at reducing thermal expansion in the aligned direction, but the thermal expansion at right angles to the fibers is not only higher than in the axial direction, but may actually exceed either of the component expansion coefficients. A further difference between αf and αm is that microscopic thermal stresses are generated, which may significantly affect the strength of the composite.

MECHANICS OF COMPOSITES

481

70

Coefficient of linear thermal expansion (10−6 K−1)

60

50

40

30

a2

20

10

0

a1

0

0.2

0.4

0.6

0.8

1.0

Volume fraction of fibers

Figure 5.88 Predicted thermal expansion coefficients in the axial and transverse direction for continuous, unidirectional glass-fiber-reinforced epoxy matrix composite. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 265. Copyright  1997 by Oxford University Press.

Finally, we turn our attention to the prediction of strength in continuous, undirectional fiber composites. Consider again the case in which a tensile stress acts parallel to the fibers (cf. Figure 5.86a). The sequence of events varies, depending on which of the two components is more brittle—that is, which extension at fracture is smaller, ∗ εf∗ or εm , where the asterisk indicates fracture (failure). Let us consider each case independently. ∗ In the case where εf∗ < εm , as in the case of carbon-reinforced epoxy, we employ the simple assumption of parallel coupling between the fibers and the matrix, for stress parallel to axis 1. We further assume that the fibers and the matrix fail independently, at the same stresses and strains as when the tensile tests are carried out on the pure materials. Our analysis begins by superimposing the fiber and matrix stress–strain

482

MECHANICS OF MATERIALS

Cooperative Learning Exercise 5.11 Recall from Section 5.1.3.1 that thermal expansion is the change in dimensions of a solid due to heating or cooling. As a result, a thermally-induced strain can result when the material is heated, as dictated by Eq. (5.29). For small deformations at constant pressure, we can thus approximate the strain as

ε=

L = αT T L

where T is the temperature change due to heating or cooling. For unidirectionally aligned fiber composites, then there will be a different strain in the axial and transverse directions for a given temperature change, since the coefficients of thermal expansion are different in the two directions. Consider, then, a composite material that consists of 60% by volume continuous, uniaxially-aligned, glass fibers in an epoxy matrix. Take the tensile modulus and Poisson’s ratio of glass to be 76 GPa and 0.22, respectively, and of the epoxy to be 2.4 GPa and 0.34, respectively. Take the coefficients of thermal expansion to be 5 × 10−6 K−1 and 60 × 10−6 K−1 for glass and epoxy, respectively. Person 1: Calculate the thermal expansion coefficient in the axial direction, α1 for this composite. Person 2: Calculate Poisson’s ratio for the composite, ν12 . Combine your information to calculate the thermal expansion coefficient in the transverse direction, α2 for this composite. The composite is now subjected to a temperature rise of 100 K. Person 1: Calculate the strain in axial direction as a result of the temperature rise. Person 2: Calculate the strain in the transverse direction as a result of the temperature rise. Compare your results. In which direction will the greatest percentage change in dimensions occur? Answers: α1 = 6.13 × 10−6 K−1 ; ν = 0.268; α2 = 34.2 × 10−6 K−1 . ε1 = 6.13 × 10−4 ; ε2 = 3.42 × 10−3 . The greatest percentage change occurs in the transverse direction, although the absolute change will depend upon the original dimensions.

curves, multiplied by their respective volume fractions, as in Figure 5.89. The total stress, σ1 , at strain, ε1 , is given by Eq. (5.84) as the sum of the two components (dashed lines), which in Figure 5.89 have already been multiplied by their respective volume fractions. The result is the solid line in Figure 5.89a and 5.89b. As strain is increased, a point is reached around εf∗ where the fibers begin to break, and their contribution to the total stress falls to zero. What happens next depends on the volume fraction of fibers, Vf . At low Vf , (Figure 5.89a), there is sufficient matrix to carry the tensile load even after all the fibers have failed. The total stress at strain ε1 is, from this point on, just the contribution from the matrix alone, (1 − Vf )σm . On further straining, the composite fails when the matrix, which now contains broken fibers, finally fails ∗ at ε1 = εm . On the other hand, at high Vf , when the fibers break, there is insufficient matrix to maintain the stress. The point of failure of the fibers then marks the failure of the composite (Figure 5.89b). In this case, the stress just prior to failure is the sum of Vf σf∗ and (1 − Vf )σm′ , where σm′ = σm εf∗ , the stress carried by the matrix at the failure strain of the fibers. So, there are two criteria for the strength at failure. At low

483

MECHANICS OF COMPOSITES

(1 − Vf )s*m

Vf s*f + (1 − Vf )s m ′

Stress

f*1

Stress

s*1

Vf sf

(1 − Vf )sm

(1 − Vf )sm

Vf sf e*f

e*m

e*f

1

1

Strain (a)

e*m

1

Strain (b)

Figure 5.89 Schematic illustration of stress–strain curves for continuous, unidirectional fiber-reinforced composites containing brittle fibers in a ductile matrix. Contributions from fibers and matrix are shown as dashed lines at (a) low fiber volume fractions and (b) high fiber volume fractions. Adapted from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 267. Copyright  1997 by Oxford University Press.

Vf and at high Vf

σ1∗ = (1 − Vf )σm∗

(5.98)

σ1∗ = Vf σf∗ + (1 − Vf )σm′

(5.99)

These two estimates for the fracture strength are shown schematically in Figure 5.90 as two intersecting straight lines. The greater of the two represents the actual strength of the composite and is indicated by a solid line. Figure 5.90 shows an important fact: A minimum fiber fraction, Vf,min , is required before the fibers exert any strengthening effect. The minimum fiber fraction can be found from Eq. (5.99): Vf,min =

σm∗ − σm′ σf∗ − σm′

(5.100)

For example, the minimum fiber fraction for a carbon-fiber-reinforced epoxy matrix composite is 0.03. ∗ In the second case where εf∗ > εm , two criteria for the failure of the composite are obtained. The analysis is similar to the previous case, but is omitted here. At low Vf σ1∗ = Vf σf′ + (1 − Vf )σm∗ and at high Vf

(5.101)

σ1∗ = Vf σf∗ (5.102) The plots of tensile strength versus volume fraction fibers corresponding to those in ∗ ∗ Figure 5.90 for εf∗ < εm are shown in Figure 5.91 for εf∗ > εm . Again, the solid line

484

MECHANICS OF MATERIALS

Strength s1*

s*f

s*m s′m 0

Vfmin

1.0 Volume fraction of fibers

Figure 5.90 Schematic illustration of tensile strength as a function of fiber volume fraction for a continuous, unidirectional fiber-reinforced, ductile matrix composite. Adapted from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 268. Copyright  1997 by Oxford University Press.

Strength s1*

s*f

s′f

s*m 0

1.0 Volume fraction of fibers

Figure 5.91 Schematic illustration of tensile strength versus fiber volume fraction for continuous, unidirectional ductile fibers in a brittle matrix. Adapted from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 269. Copyright  1997 by Oxford University Press.

MECHANICS OF COMPOSITES

485

represents the behavior of the composite. Equations (5.98)–(5.99) and (5.101)–(5.102) have been found to predict well the experimental tensile strengths of brittle fiber/ductile matrix and ductile fiber/brittle matrix composites at a variety of fiber volume fractions in undirectional, continuous-fiber-reinforced composites. 5.4.2.2 Fiber End Effects*. When considering discontinuous fibers, it is necessary to take into account the fiber ends, which are weak points in the composite. Here stresses tend to concentrate. The critical issue is how stresses get transferred from the low-strength, low-modulus matrix to the high-strength, high-modulus fiber reinforcement. Consider an isolated short fiber in a polymer matrix under tension, as illustrated in Figure 5.92. It must be assumed that negligible stress gets transferred to the fibers across their end faces. It is the shear stress at the fiber–matrix interface, τi , that performs the vital task of transmitting tension to the fiber. It can be shown that the relationship between the axial tension in the fiber, σf , and the interfacial shear stress is −4τi dσf = (5.103) dx d

where d is the fiber diameter and dσf /dx is the stress distribution along the fiber axis direction, x. Predicting σf and τi involves stress analysis of the fiber–matrix composite under load. A simplified model assumes perfect bonding between the fibers and the matrix, and it results in the following prediction of σf as a function of distance x, along the fiber, measured from its center:   na2x   cosh   l   σ f = E f εf 1 −   cosh(na)         





(5.104)

l

d (a)

(b)

Figure 5.92 Schematic illustration of fiber in polymer matrix (a) in the undeformed state and (b) under a tensile load. Horizontal lines are shown to demonstrate strain distribution. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 242. Copyright  1997 by Oxford University Press.

486

MECHANICS OF MATERIALS

where a is the fiber aspect ratio (length to diameter, l/d) and n represents a dimensionless group of constants: 2Gm (5.105) n= Ef ln(2R/d) where 2R is the distance from the fiber to its nearest neighbor, d is again the fiber diameter, and Gm and Ef are the shear modulus and elastic modulus of the matrix and fiber, respectively. Equation (5.104) can be substituted into Eq. (5.103) and differentiated to obtain the interfacial shear stress: 

na2x sinh nEf εf l τi = 2 cosh(na)



(5.106)

Equation (5.106) shows that the critical parameter dictating how rapidly stress builds up along the fiber is the product na. This is illustrated in Figure 5.93, where graphs of σf and τi /n are plotted versus x/ l for three values of na. These plots show that for the most efficient stress transfer to the fibers, the product na should be as high as possible. This confirms the desirability of a high aspect ratio, but also shows that n should be high. The ratio of moduli, Gm /Ef , should therefore be as high as possible. Typical values encountered in real composites are a = 50, n = 0.24, so na = 12 for 30 vol% glass fibers in a nylon matrix.

x

Interface shear stress/n ti /n (units of Ef e1)

Fiber tensile stress sf (units of Ef e1)

−l /2

0

l /2

na = 2

5

10

na = 2

5

10

1

0

0.5 0 −0.5

Figure 5.93 Distribution of tensile stress (top) and interfacial shear stress (bottom) along a short fiber. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 276. Copyright  1997 by Oxford University Press.

MECHANICS OF COMPOSITES

s1

487

s1

Figure 5.94 Schematic illustration of aligned discontinuous fibers in polymer matrix. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 277. Copyright  1997 by Oxford University Press.

The most important effects of the fiber tensile stress falling away to zero at the ends are reductions in the axial tensile modulus and strength of the composite. Consider the idealized composite shown in Figure 5.94, where the short fibers are in parallel, and an imaginary (dash) line is drawn across the composite at right angles to the fibers. It can be shown that the stress carried by the composite, σ1 , is given by σ1 = Vf σ f + (1 − Vf )σm

(5.107)

where Vf is the fiber volume fraction, σm is the matrix stress, and σ f is the mean fiber stress, given by

1 l/2 σf = σf dx (5.108) l −l/2 Substitution of Eq. (5.104) into (5.108) and integration yields σ f = E f ε1



tan h(na) 1− na



= E f ε 1 β1

(5.109)

The axial tensile modulus, E1 , is obtained by dividing the stress σ1 by the strain ε1 : E1 = β1 Vf Ef + (1 − Vf )Em

(5.110)

Note the similarity of this relationship to the equation of mixing given at the beginning of this section by Eq. (5.82). The factor β1 is called the fiber length correction factor, and it corrects the fiber modulus for the shortness of the fibers. It is plotted in Figure 5.95 as a function of the product na. When na becomes very large, β1 approaches one, as is expected since this limit is the case of continuous fibers, and Eq. (5.110) reduces to Eq. (5.82). When na falls below about 10, β1 is significantly less than one. As indicated in Figure 5.93, the interfacial shear stress, τi , is concentrated at the fiber ends. With increasing strain, these are the locations where the interface first fails and debonding of the matrix from the fiber begins. This occurs when τi reaches the interfacial shear strength, the magnitude of which is determined by three factors: the strength of the chemical bond between the fibers and the matrix; the friction between the fibers and the matrix resulting from matrix pressure introduced during cooling;

488

MECHANICS OF MATERIALS

Fibre length correction factor

1.0

0.8

0.6

0.4

0.2

0 0.1

1

10

100

na (log scale)

Figure 5.95 Fiber length correction factor for modulus, β1 , as a function of the geometric factor, na. Adapted from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 277. Copyright  1997 by Oxford University Press.

and the shear strength of the matrix. With further straining, the debonding regions of the fiber slip within the matrix. Slippage is resisted by a constant friction stress, τi∗ , and the stress distributions switch from those shown in Figure 5.93 to those shown in Figure 5.96. Since the fiber ends now carry a constant interfacial shear stress, the tensile stress has a constant gradient, given by Eq. (5.103), and stress builds up linearly from each end. To a good approximation, only the stresses shown in the shaded area of Figure 5.96 need to be considered. The length of the debonded region, δ, can be found by applying Eq. (5.103) to either end of the fiber and rearranging: δ=

E f ε1 d 4τi∗

(5.111)

As straining proceeds, the debonded region grows with increasing strain. At the point of fiber breakage, the strain is ε1 = εf∗ = σf∗ /Ef , and the debonded regions have lengthened to: σf∗ d δ∗ = (5.112) 4τi∗ However, the debonded region can only reach this value if the fibers are longer than a critical length, lc , which is the sum of the debonded regions at each end of the fiber: lc = 2δ ∗ =

σf∗ d 2τi∗

(5.113)

If the fibers are shorter than this critical length, failure of the composite occurs after the debonded regions have extended along the full length of the fiber. The entire

MECHANICS OF COMPOSITES

x

d

489

d

Interface shear stress ti

Fiber tensile stress sf

0 Debonding

_ Ef e 1 ∼ 0 ti* 0 − ti*

Figure 5.96 Distribution of fiber tensile stress and interfacial shear stress after fiber–matrix debonding has commenced. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 280. Copyright  1997 by Oxford University Press.

fiber is then merely slipping within the matrix and it cannot be loaded to its failure stress; the fiber is not properly performing its reinforcing function. As a result, lc is an important parameter, and efforts are made to produce composites in which the fiber length exceeds lc . The axial strength of the composite can be calculated by applying Eq. (5.107). In the most common case, the matrix is more ductile than the fibers, such that the matrix ∗ critical strain exceeds the fiber critical strain, εm > εf∗ . At low volume fraction of fibers, the fibers located with their ends within a distance δ ∗ of the failure plane do not fracture, but rather simply pull out of the matrix. This movement is resisted by the friction stress, τi∗ . The proportion of fibers for which this occurs is lc / l, and the mean stress resisting pull-out is σf∗ /2. The fibers therefore contribute a mean stress of σ f = σf∗ lc /2l and from Eq. (5.107) the strength becomes σ1∗ =

lc Vf σf∗ + (1 − Vf )σm∗ 2l

(5.114)

At high volume fractions of fibers, the mean fiber stress at failure is obtained from the trapezoidal-shaped distribution of tensile stress illustrated in Figure 5.97, with ε1 = σf∗ /Ef and δ = δ ∗ = lc /2, such that   lc σ f = σf∗ 1 − 2l

(5.115)

490

MECHANICS OF MATERIALS

Tensile strength s*1

s*f

s*f 1−

s*f

lc 2l

lc 2l

s*m s′m 0

1.0 Volume fraction of fibers

Figure 5.97 Plot of composite tensile strength versus fiber volume fraction for an aligned, short fiber-reinforced polymer matrix composite. The dotted lines show the corresponding values for continuous fibers for comparison purposes. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 281. Copyright  1997 by Oxford University Press.

and the strength is again given by Eq. (5.107):   lc ∗ Vf σf∗ + (1 − Vf )σm′ σ1 = 1 − 2l

(5.116)

where σm′ indicates the stress carried by the matrix at the failure strain of the fibers, σm εf∗ . The equations for the composite tensile strength in the two extremes of low volume fraction fibers and high volume fraction fibers, Eqs. (5.114) and (5.116), respectively, are plotted as a function of volume fraction of fibers in Figure 5.97. Here, the dashed lines represent the low Vf and high Vf limits, with the greater of the two values shown as a solid line. The solid line represents the predicted strength of the composite. For comparison purposes, the corresponding strength equations at low- and high-volume fractions of fibers are shown as dotted lines for continuous-fiber composites. Figure 5.97 clearly shows that reducing the fiber length has two major effects on the failure of composites—namely, that the strength is reduced (except at very low volume fractions) and there is a wider range of volume fractions over which failure is matrix-controlled. 5.4.2.3 Composite Moduli: Halpin–Tsai Equations∗ . Derivations of estimates for the effective moduli (tensile E, bulk K, and shear G) of discontinuous-fiberreinforced composite materials are extremely complex. The basic difficulty lies in the complex, and often undefined, internal geometry of the composite. The problem has been approached in a number of ways, but there are three widely recognized

MECHANICS OF COMPOSITES

491

Cooperative Learning Exercise 5.12

An injection molded bar contains 20% by volume of short carbon fibers in a matrix of nylon 6,6. Take the tensile strengths of the carbon fibers and the nylon to be 3200 MPa and 70 MPa, respectively, and assume the shear strength of the carbon fiber-nylon interface to be 32 MPa. The tensile moduli of nylon 6,6 and carbon fibers are 2.7 GPa and 230 GPa, respectively. Assume that the fibers are all of length 400 µm and diameter 6 µm and are perfectly aligned along the axis of the bar. ∗ , from the matrix properties. Person 1: Calculate the strain at failure in the matrix, εm ∗ Person 2: Calculate the strain at failure in the fiber, εf , from the fiber properties. Compare your answers. Will the failure occur in the matrix or the fiber first? Person 1: Calculate the matrix stress at failure based on which component fails first. Person 2: Calculate fiber critical length lc . Combine your information to calculate the composite tensile strength assuming that Vf is a high value. Equation (5.96), σ1∗ = (1 − 300/800)(0.2)(3200 MPa) + (1 − 0.2)(37.5 MPa) = 430 MPa σm′ = Em εf∗ = 2700(0.0139) = 37.5 MPa;

lc =

2τi∗ σf∗ d

=

(3200)(6) = 300 µm 2(32)

∗ = σm∗ /Em = 70/2700 = 0.026; εf∗ = σf∗ /Ef = 3200/230000 = 0.0139 Answers: εm ∗ , fiber fails prior to matrix, so strain at failure is 0.0139. εf∗ < εm

(Adapted from McCrum, p. 282)

approaches: bounding methods, a special model called the composite cylinder assemblage, and semiempirical equations called the Halpin–Tsai equations. In all cases, only unidirectional, discontinuous fibers are considered—that is, short fibers that are preferentially aligned along a certain axis. Two directions are important in this analysis, as defined in Figure 5.98: the longitudinal direction, along the direction of the applied force, and the transverse direction, perpendicular to the applied force. Only the bounding methods and Halpin–Tsai approximations will be described here. Upper and lower bounds on the elastic constants of transversely isotropic unidirectional composites involve only the elastic constants of the two phases and the fiber volume fraction, Vf . The following symbols and conventions are used in expressions for mechanical properties: Superscript plus and minus signs denote upper and lower bounds, and subscripts f and m indicate fiber and matrix properties, as previously. Upper and lower bounds on the composite axial tensile modulus, Ea , are given by the following expressions: Ea+ = Vf Ef + (1 − Vf )Em + Ea− = Vf Ef + (1 − Vf )Em +

4Vf (1 − Vf )(νf − νm )2 Vf /Km + (1 − Vf )/Kf + 1/Gf

4Vf (1 − Vf )(νf − νm )2 Vf /Km + (1 − Vf )/Kf + 1/Gm

(5.117) (5.118)

Where ν is Poisson’s ratio, K is the bulk modulus, and G is the shear modulus. Similar relations exist for Poisson’s ratio, bulk modulus, and shear moduli in both the axial

492

MECHANICS OF MATERIALS

Longitudinal direction

Transverse direction

Figure 5.98 Schematic illustration of discontinuous, aligned fiber-reinforced composite.

and transverse directions [8]. Note that the final terms in both expressions provide the corrections to the simple rule-of-mixture approximation. For most materials, the bounds on axial tensile modulus tend to be reasonably close together, and the rule-of-mixtures prediction given by Eq. (5.82) is generally accurate enough for practical purposes for both the elastic modulus Ea = Vf Ef + (1 − Vf )Em

(5.119)

and, to a lesser extent, Poisson’s ratio νa = Vf νf + (1 − Vf )νm

(5.120)

The rule of mixtures generally provides a poor estimate of the bulk and shear moduli in both the axial and transverse directions, respectively, in unidirectional, discontinuousfiber-reinforced composites. The Halpin–Tsai equations represent a semiempirical approach to the problem of the significant separation between the upper and lower bounds of elastic properties observed when the fiber and matrix elastic constants differ significantly. The equations employ the rule-of-mixture approximations for axial elastic modulus and Poisson’s ratio [Equations. (5.119) and (5.120), respectively]. The expressions for the transverse elastic modulus, Et , and the axial and transverse shear moduli, Ga and Gt , are assumed to be of the general form   1 + Vf ξ η (5.121) P = Pm 1 − Vf η where P is the composite property (modulus) of interest, Pm is the corresponding matrix property, Vf is the volume fraction of fibers, and ξ and η are correlations

MECHANICS OF COMPOSITES

as follows:

 Pf −1 P  η=  m Pf +ξ Pm 

493

(5.122)

where Pf and Pm are again the fiber and matrix properties of interest, respectively, and ξ = l/d

(5.123)

where l/d is the fiber aspect ratio. 5.4.2.4 Discontinuous-Fiber-Reinforced Polymer–Matrix Composites: Sheet Molding Compound∗ . Of the parameters influencing the mechanical properties in short-fiber-reinforced polymer–matrix composites, fiber composition, matrix composition, fiber geometry, and manufacturing method will be elaborated upon here. The composition and properties of some common reinforcing fibers were given earlier in Table 1.31, the most common of which is glass fiber. The compositions of some commercial glass fibers were given earlier in Table 1.32, and the most common polymer matrices were given in Table 1.28. For this section, we will concentrate on discontinuous-glass-fiber-reinforced polymer–matrix composites. Glass fiber loadings from 10% to 40% by weight produce increases in strength and rigidity, along with a marked decrease in the coefficient of thermal expansion. Some of these properties develop gradually with increasing glass fiber content, while others change sharply. Both modulus and tensile strength are proportional to glass content. Figure 5.99 shows these relationships for short glass-fiber-reinforced polypropylene and styrene–acrylonitrile copolymer. Glass fiber reinforcement often improves impact strength, but long-fiber reinforcements are more effective in this regard than short-fiber composites. The most significant effect of glass-fiber reinforcement in thermoplastics is the retention of impact strength down to very low temperatures, but the relationship between glass content and impact strength is not always linear. With low-modulus thermoplastics, optimum impact strength may or may not be achieved at less than the maximum glass content. Normally, room temperature impact strengths of low elastic modulus materials suffer by incorporating glass reinforcement. With rigid thermoplastics, marked improvement in room-temperature impact strengths usually occur with increasing glass content. In virtually all thermoplastics, impact strengths at low temperatures improve with increasing glass content. The reinforcement can also cause increases in hardness and abrasion resistance, along with decreases in creep and dimensional changes with humidity. Undesirable effects brought on by the addition of glass reinforcement include opacity, low surface gloss, some difficulty in electroplating, and loss of mechanical flexibility. The effect of fiber diameter on the tensile strength of a glass-fiber-reinforced polystyrene composite is shown in Figure 5.100. Some reinforcements also have a distribution of fiber diameters that can affect properties. Recall from the previous section that the fiber aspect ratio (length/diameter) is an important parameter in some mechanical property correlations.

494

MECHANICS OF MATERIALS

18

1.8

16

1.6

14

1.4

Flexural Modulus, 106 psi

Tensle Strength, 1000 psi

2.0

12 SAN-Glass 10 8 6

1.2 SAN-Glass 1.0 0.8 0.6

PP-Glass 4

0.4

2

0.2

0

0

10 20 30 40 Glass Content, %

0

50

PP-Glass

0

10 20 30 40 Glass Content, %

50

(b)

(a)

Figure 5.99 Relationship of glass fiber content with (a) tensile strength and (b) flexural modulus for styrene acrylonitrile (SAN) and polypropylene (PP). Reprinted, by permission, from G. Lubin, Handbook of Fiberglass and Advanced Plastics Composites, p. 130. Copyright  1969 by Van Nostrand Reinhold.

Tensile Strength, 1000 psi

13

12

11

10

9

0

20

40

60

80 −5

Fiber Diameter, 10

100

120

in.

Figure 5.100 Effect of glass fiber diameter on strength of reinforced polystyrene. Reprinted, by permission, from G. Lubin, Handbook of Fiberglass and Advanced Plastics Composites, p. 130. Copyright  1969 by Van Nostrand Reinhold.

MECHANICS OF COMPOSITES

495

A number of processing methods exist to produce glass-fiber/polymer–matrix composites, including reinforced reaction injection molding (RRIM), resin transfer molding (RTM), filament winding, and compression molding of sheet molding compound (SMC), to name a few. These processing techniques will be elaborated upon in Chapter 7. For this section, we concentrate on one processing technique that is commonly used to make useful components out of discontinuous-glass-fiberreinforced polymer matrix composites—namely, compression molding of SMCs. This method comprises a wide range of thermosetting prepreg molding materials reinforced generally with short glass fiber strands about 25–50 mm long. Traditionally, unsaturated polyester resins have been used for SMC; however, compositions based on vinyl ester, epoxy, and urethane–polyester hybrids are currently in use. The most commonly used fiber in sheet molding compounds is E-glass (see Table 1.32), although other fibers such as S-2 glass, carbon, and Kevlar 49 are also used. Both short (discontinuous) and long (continuous) forms of fibers are used in SMC. Short fibers are commonly used in applications requiring isotropic properties. Long fibers, which will be described in the next section, provide high strength and modulus for structural applications. The conventional SMC material uses a continuous fiber roving, which is a collection of 25–40 untwisted fiber strands. Each fiber strand typically contains 200–400 E-glass filaments. Short fibers are obtained by chopping the rovings into desired lengths at the time of SMC sheet production. Sheet molding compounds are designated according to the form of fibers (see Figure 5.101). In both SMC-R and SMC-CR, the fiber content is shown by the weight percent at the end of each letter designation; for example, SMC-R40 contains 40 wt% of randomly oriented short fibers. For the remainder of this section, we will concentrate on randomly oriented short-fiber sheet molding compound (SMC-R) in order to illustrate some of the common mechanical properties of discontinuous-glass-fiber-reinforced polymer-matrix composites. The tensile stress–strain diagram for two SMC-R compounds are shown in Figure 5.102. Unlike the unreinforced polymers, SMC composites do not exhibit any yielding and their strain to failure is low, usually less than 2%. However, at low fiber contents, the tensile stress–strain diagrams are nonlinear, which indicates the presence of microdamage, including crazing. At high fiber contents, the nonlinearity decreases, but the stress–strain diagram exhibits a bilinearity with a break point (knee) close to the matrix failure strain. At stress levels higher than the break point, the slope of the

L

L

T

SMC-R (a)

SMC-CR (b)

T

XMC (c)

Figure 5.101 Various forms of sheet molding compounds: (a) random, SMC-R; (b) continuous/ random, SMC-CR; and (c) criss-crossed continuous/random XMC; L and T refer to longitudinal and transverse directions, respectively. Reprinted, by permission, from Composite Materials Technology, P. K. Mallick and S. Newman, eds., p. 33. Copyright  1990 by Carl Hanser Verlag.

496

MECHANICS OF MATERIALS

200

SMC-R65

STRESS (MPa)

150

100 SMC-R25 50

0

0

.5

1.0

1.5

2.0

STRAIN (%)

Figure 5.102 Tensile stress–strain diagrams for two SMC compounds: 25 wt% fiber, bottom, and 65 wt% fiber, top. Reprinted, by permission, from Composite Materials Technology, P. K. Mallick and S. Newman, eds., p. 45. Copyright  1990 by Carl Hanser Verlag.

stress–strain diagram is lower than the initial slope. The tensile strengths of SMCR composites increase significantly with increasing fiber content, but the modulus is affected only moderately. The following empirical relationships for tensile strength, σt , and tensile modulus, E, of SMC composites have been developed: σt = 0.33Vf σf + 0.31(1 − Vf )σm

(5.124)

E = 0.59Vf Ef + 0.71(1 − Vf )Em

(5.125)

and

where σf and σm are the ultimate tensile strengths of the fiber and matrix, respectively, Ef and Em are their respectively tensile moduli, and Vf is the volume fraction of fibers. Note the similarity in form of these equations to the rule of mixtures given by Eq. (5.82). The tensile strength of SMC-R composites increases as the chopped fiber length is increased, but fiber length has virtually no effect on the tensile modulus. The use of S-glass instead of E-glass (see Table 1.32) has a significant effect on increasing the mechanical properties of the composite. The resin type can also influence the tensile properties, particularly at low fiber contents. The temperature dependence of both tensile strength and modulus for a SMC-R50 composite are shown in Figure 5.103. Similar effects are also seen for the flexural strength and modulus. As with unreinforced polymers, humidity and liquids environments can affect strength and modulus. Flexural strengths tend to be higher than tensile strengths in SMC composites. Flexural load-deflection modulus values are nonlinear, which indicates the occurrence of microcracking even at low loading. In general, flexural properties follow the same trends as the tensile properties and are affected by fiber content, fiber lengths, type, and orientation.

497

MECHANICS OF COMPOSITES

58.0

25

43.5 15 29.0

Strength (MPa)

Modulus (GPa)

20

10

5 −50

−25

0

25

50

75

100

14.5 125

Temperature (°C)

Figure 5.103 Temperature dependence of tensile strength and modulus for SMC-R50 composite. Reprinted, by permission, from P. K. Mallick and S. Newman, eds., Composite Materials Technology, p. 59. Copyright  1990 by Carl Hanser Verlag.

Table 5.8 Selected Compressive and Shear Properties of Various SMC Composites Property Compressive strength, MPa Compressive modulus, GPa In-plane shear strength, MPa In-plane shear modulus, GPa Interlaminar shear strength, MPa

SMC-R25

SMC-R50

SMC-R65

183 11.7 79 4.48 30

225 15.9 62 5.94 25

241 17.9 128 5.38 45

Source: Composite Materials Technology, P. K. Mallick and S. Newman, ed. Copyright  1990 by Carl Hanser Verlag.

As with tensile properties, both compressive strength and modulus depend on the fiber content and fiber orientation (see Table 5.8). The interlaminar shear strength reported in Table 5.8 is a measure of the shear strength in the thickness direction of the SMC sheet. It is determined by three-point flexural testing of beams with short spanto-depth ratios and is considered to be a quality-control test for molded composites. Most of the data available on the fatigue properties of SMC composites are based on tension-tension cyclic loading of unnotched specimens. The S –N diagram (cf. Section 5.2.1.4) of an SMC-R65 composite at various temperatures is shown in Figure 5.104. In general, the fatigue strength of SMC-R composites increases with increasing fiber content. The microscopic fatigue damage appears in the forms of matrix cracking and fiber–matrix interfacial debonding. The matrix cracking is influenced by the stress concentrations at the chopped fiber ends, as described in Section 5.4.2.1. Matrix cracks are generally normal to the loading direction. However, in fiber-rich areas, with fibers oriented parallel to the loading direction, the matrix crack lengths are small and limited by the interfiber spacing.

498

MECHANICS OF MATERIALS

Maximum stress amplitude (MPa)

300 250

R = 0.05 5 Hz

200

150 100 23 °C −40 °C 93 °C

50 0 0.1

1

10

102 103 Number of cycles (N)

104

105

106

Figure 5.104 Fatigue S –N diagram for SMC-R65 composite. Reprinted, by permission, from Composite Materials Technology, P. K. Mallick and S. Newman, eds., p. 52. Copyright  1990 by Carl Hanser Verlag. 7000 Stress = 44.9 MPa 100°C

Strain (10−6)

6000

70 °C 5000 40 °C 27 °C 4000

0

2

4

6

8

10

Time (103 min)

Figure 5.105 Creep response of SMC-R50 composite at various temperatures. Reprinted, by permission, from P. K. Mallick and S. Newman, eds., Composite Materials Technology, p. 62. Copyright  1990 by Carl Hanser Verlag.

A power-law expression of the form of Eq. (5.115) adequately describes the creep properties of SMC composites. The creep response of the SMC-R50 composite at various temperatures shown in Figure 5.105 is representative of this behavior. Among the material variables, fiber content once again has the greatest influence on the creep strain. At a given temperature and stress level, creep strain is higher if the fiber content is reduced.

499

MECHANICS OF COMPOSITES

The mechanical behavior of other discontinuous-glass-fiber-reinforced polymer–matrix composites, while highly variable, exhibits many of the same functional dependences as SMC composites. 5.4.2.5 High-Performance Polymer–Matrix Composites∗ . We continue our survey of important fiber–matrix composites in this section by extending the principles introduced in the previous section to continuous fibers. Certainly the continuous-fiber versions of sheet molding compounds (SMCs) of the previous section (see Figure 5.101) can be elaborated upon here, but we shall take a slightly different approach. Just as the primary consideration in the use of discontinuous fibers for SMCs was low cost, the primary purpose in utilizing continuous fibers over discontinuous fibers is to improve mechanical properties. It makes sense, then, to take this benefit to its extreme and describe continuous-fiber-reinforced composites with the ultimate in mechanical properties. Typically, these are not reinforced with glass fibers, but rather with so-called high-performance fibers such as Kevlar , boron, and carbon. The specific mechanical properties of these fibers are compared in Figure 5.106. Notice the enormous improvement in both strength and modulus relative to E-glass, and even S-glass fibers. We will concentrate on carbon fibers for illustrative purposes, but also present data on Kevlar and other types of continuous fibers where appropriate. Polymer–matrix composites will still be the focus of our descriptions, though, since it is the lightweight matrix that provides the added benefit of weight savings to the composite. The combination of 106 cm 10

5 KEVLAR 29

15

KEVLAR 49 25

RESIN-IMPREGNATED STRANDS 20

8 S-GLASS

HT GRAPHITE

6

15 BORON

OTHER ORGANICS

4

10

HM GRAPHITE

E-GLASS STEEL

2

106 cm

SPECIFIC TENSILE STRENGTH, 106 in

10

5

ALUMINUM 0

0

1

2

3

4

5

6

7

SPECIFIC TENSILE MODULUS, 106 in.

Figure 5.106 Specific tensile strength and specific tensile modulus of reinforcing fibers. Reprinted, by permission, from M. Schwartz, Composite Materials Handbook, 2nd ed., p. 2.70. Copyright  1992 by McGraw-Hill, Inc.

500

MECHANICS OF MATERIALS

mechanical properties and weight savings that carbon-fiber-reinforced polymer (CFRP) composites offer is phenomenal. For example, if a civilian aircraft were constructed, where possible, from CFRP instead of aluminum alloy, the total weight reduction would be approximately 40%. This creates enormous potential for increased fuel efficiency and increased payload. Recall from Section 1.4.5.1 that there are two primary types of carbon fibers: polyacrylonitrile (PAN)-based and pitch-based. There are also different structural forms of these fibers, such as amorphous carbon and crystalline (graphite) fibers. Typically, PANbased carbon fibers are 93–95% carbon, whereas graphite fibers are usually 99+%, although the terms carbon and graphite are often used interchangeably. We will not try to burden ourselves with too many distinctions here, since the point is to simply introduce the relative benefits of continuous-fiber composites over other types of composites, and not to investigate the minute differences between the various types of carbon-fiber-based composites. The interested reader is referred to the abundance of literature on carbon-fiber-reinforced composites to discern these differences. The primary resin of interest is epoxy. Carbon-fiber–epoxy composites represent about 90% of CFRP production. The attractions of epoxy resins are that they polymerize without the generation of condensation products that can cause porosity, they exhibit little volumetric shrinkage during cure which reduces internal stresses, and they are resistant to most chemical environments. Other matrix resins of interest for carbon fibers include the thermosetting phenolics, polyimides, and polybismaleimides, as well as high-temperature thermoplastics such as polyether ether ketone (PEEK), polyethersulfone (PES), and polyphenylene sulfide. Unlike short-fiber composites, modulus and strength in long-fiber composites are little affected by the interface. However, the interface does affect off-axis properties, such as compressive modulus, and the toughness and shear strength are affected. As shown in Figure 5.107, a reduction in the interfacial strength, such as through the

Work of Fracture (kJm−2)

50

40

silicone oil

untreated etched silane-coated

30

20

10

commercial coating impact slow bend

20

40

60

Shear Strength (MPa)

Figure 5.107 Variation in toughness and shear strength for various surface treatments of continuous carbon fiber-reinforced composites. Reprinted, by permission, from T. L. Vigo and B. J. Kinzig, ed., Composite Applications, p. 224. Copyright  1992 by VCH Publishing, Inc.

MECHANICS OF COMPOSITES

501

treatment of the fiber surface with a coating, increases toughness, represented here as the work of fracture, while simultaneously reducing the shear strength. We will describe the mechanism behind this toughening effect in Section 5.4.2.7. Due to the fact that the mechanical properties of unidirectional, continuous-fiberreinforced composites are highly anisotropic, maximum effectiveness is often achieved by making laminate composites of multiple layers. This is particularly true of carbon and Kevlar -reinforced polymers, which will be described in Section 5.4.3. 5.4.2.6 Whiskers and Metal–Matrix Composites∗ . Metal–matrix composites (MMCs) are of interest because they offer a combination of properties not available through alloying or processing methods alone. Specifically, light-metal MMCs with aluminum, titanium, or magnesium matrices have very high specific strengths and specific moduli relative to alloys, owing to the high-tensile-strength and high-modulus fibers that are incorporated in them. The plastic deformation of the metal–matrix composite imparts excellent toughness to the MMC. The common types of metals used in MMCs were first introduced in Table 1.29, and the types of reinforcing fibers were first listed in Table 1.31. Many of the principles previously discussed apply to MMCs: discontinuous versus continuous reinforcement, effect of fiber geometry (aspect ratio), the effect of interfacial bonding on composite properties, and variations due to processing methods. Just as we focused on glass fibers in discontinuous polymer–matrix composites, and carbon fibers for continuous polymer–matrix composites, we will select one system as representative of MMCs. Certainly, carbon fibers find widespread use in MMCs, as do boron fibers, alumina (Al2 O3 ) fibers, and Kevlar . However, the most widely varied usages employ SiC as a reinforcement in MMCs. The selection of SiC as the reinforcement of choice in this section also allows us to discuss a topic that has not yet been addressed: whisker technology. Though the distinction between a whisker and a fiber is sometimes vague, whiskers are generally considered a subclass of fibers that have been grown in controlled conditions, often in the vapor phase, which leads to high-purity single crystals. The resultant highly ordered structure produces not only unusually high strengths, but also significant changes in electrical, optical, magnetic, and even superconductive properties. As the diameter of the whisker is reduced, its strength increases rapidly due to the increasing perfection in the fiber. The tensile strengths of whiskers are far above those of the more common, high-volume production reinforcements. This is illustrated in Figure 5.108, where the schematic stress–strain curves of whiskers and other reinforcements are compared. Notice that whiskers can be strained elastically as much as 3% without permanent deformation, compared with less than 0.1% for bulk ceramics. Whiskers have been produced in a range of fiber sizes and in three forms: grown wool, loose fiber, and felted paper. The wool has a fiber diameter of 1–30 µm and aspect ratios of 500–5000. The wool bulk density is about 0.03 g/cm3 , which is less than 1% solids. This is a very open structure, suitable for a vapor-deposition coating on the whiskers or direct use as an insulation sheet. Loose fibers are produced by processing the larger diameter fiber to yield lightly interlocked clusters of fibers with aspect ratios of 10–200. In felt or paper, the whiskers are randomly oriented in the plane of the felt and have fiber aspect ratios ranging from 250 to 2500. The felted paper has approximately 97% void volume and a density of 0.06–0.13 g/cm3 . Whiskers can be formed from a variety of materials, as illustrated in Table 5.9, but there are only two general methods for forming them: basal growth and tip growth.

502

MECHANICS OF MATERIALS

Tensile strength (GPa)

21

14 whiskers

7 HMG

0

boron

10

Kevlar ®

20

glass

30

Elongation (%)

Figure 5.108 Schematic comparison of stress–strain diagrams for common reinforcing fibers (HMG = high modulus graphite) and whiskers. Reprinted, by permission, from A. Kelly, ed., Concise Encyclopedia of Composite Materials, revised edition, p. 312. Copyright  1994 by Elsevier Science Publishers, Ltd. Table 5.9

Selected Mechanical Properties of Some Common Whiskers

Material Aluminum oxide Aluminum nitride Beryllium oxide Boron oxide Graphite Magnesium oxide Silicon carbide (α) Silicon carbide (β) Silicon nitride

Density (103 kg m−3 )

Melting point (◦ C)

Tensile Strength (GPa)

Young’s modulus (GPa)

3.9 3.3 1.8 2.5 2.25 3.6 3.15 3.15 3.2

2082 2198 2549 2449 3593 2799 2316 2316 1899

14–28 14–21 14–21 7 21 7–14 7–35 7–35 3–11

550 335 700 450 980 310 485 620 380

Source: Concise Encyclopedia of Composite Materials, A. Kelly, ed., revised edition. Copyright  1994 by Elsevier Science Ltd.

In basal growth, the atoms of growth migrate to the base of the whisker and extrude the whisker from the substrate. Tip growth involves fabrication at high temperatures such that the vapor pressure of the whisker-forming material becomes significant to the point that the atoms attach themselves to the tip of the growing whisker. Silicon carbide whiskers are generally formed by the pyrolysis of chlorosilanes in hydrogen at temperatures above 1400◦ C. Selection of proper silanes, optimum reaction conditions, and suitable reaction chambers are important process parameters. In some cases, a coating is also produced on the whiskers to reduce reaction between the whisker and the metal matrix after composite formation.

MECHANICS OF COMPOSITES

503

Whiskers can be incorporated into the metallic matrix using a number of compositeprocessing techniques. Melt infiltration is a common technique used for the production of SiC whisker–aluminum matrix MMCs. In one version of the infiltration technique, the whiskers are blended with binders to form a thick slurry, which is poured into a cavity and vacuum-molded to form a pre-impregnation body, or pre-preg, of the desired shape. The cured slurry is then fired at elevated temperature to remove moisture and binders. After firing, the preform consists of a partially bonded collection of interlocked whiskers that have a very open structure that is ideal for molten metal penetration. The whisker preform is heated to promote easy metal flow, or infiltration, which is usually performed at low pressures. The infiltration process can be done in air, but is usually performed in vacuum. The benefits of SiC whisker-reinforced aluminum can be seen in Figure 5.109. The room-temperature stiffness (modulus) of the composite is about 50% higher than that Ultimate tensile strength (MPa) 0

200

400

600

At room temperature

At 250°C

Elastic modulus (GPa) 0

50

100

150

At room temperature

Unreinforced 6061 aluminum

20 vol % SiC whiskers on 6061 aluminum

Figure 5.109 Variation of tensile strength and tensile modulus with temperature of SiC-reinforced aluminum. Reprinted, by permission, from A. Kelly, ed., Concise Encyclopedia of Composite Materials, revised edition, p. 189. Copyright  1994 by Elsevier Science Publishers, Ltd.

504

MECHANICS OF MATERIALS

TEST TEMPERATURE, °F 300

500

700

900

1100

1300

1500 175

Ti (A-70) 36 vol % SiC Ti (A-75) BAR Ti (A-70) PRESSED FOIL Ti (6-4) BAR

1000 800

150 125 100

600 75 400

50

200 0 37.8

25 232

427

621

TENSILE STRENGTH, kpsi/in2

TENSILE STRENGTH, MPa

100 1200

0 816

TEST TEMPERATURE (°C)

Figure 5.110 Variation of tensile strength with temperature for some titanium-based alloys and composites. Reprinted, by permission, from M. Schwartz, Composite Materials Handbook, 2nd ed., p. 2.111. Copyright  1992 by McGraw-Hill, Inc.

of the aluminum alloy. When the lower density of aluminum is considered, the specific modulus is almost twice that of pure titanium. The primary benefit, however, is in the high-temperature strength retention of the composite relative to the unreinforced aluminum. This effect is also illustrated in Figure 5.110 for SiC fiber-reinforced titanium matrix composite relative to the base metal formed by different techniques. Unlike dispersion-strengthened metals (cf. Section 5.4.1.1), where the particles strengthen by blocking the movement of dislocations in the matrix, fiber reinforcements improve strength by a transference of load from the matrix to the fiber phase. Thus, the relationships for composite strength as a function of volume fraction fiber [Eqs. (5.114) and (5.116)] generally apply to metal-matrix composites, as does Figure 5.97. The ′ primary difference is that the matrix strength at fiber failure, σm , is subject to strain hardening in the metallic matrix. With regard to composite modulus, the contribution of the ductile metal matrix to the composite stiffness is complex. A complete description of the internal state of stress for every applied load is required, and it is being approached with analytical and finite-element models. Non-whisker SiC fibers are also of importance in MMCs, and they are currently available in two commercial forms: Tyranno and Nicalon . As with the whisker form, the primary advantages of SiC fibers is their oxidation resistance and high-temperature mechanical property retention relative to other fibers. The high-temperature strength of three commercially available Nicalon SiC fibers is shown in Figure 5.111. In addition to aluminum, other types of metal matrices in MMCs include magnesium, which is relatively easy to fabricate due to its low melting point, lead for batteries, titanium for aircraft turbine engines, copper for magnetohydrodynamics, and iron, nickel, or cobalt alloys.

MECHANICS OF COMPOSITES

505

1.6

1.4

Applied stress s (GPa)

1.2 1 0.8 0.6 0.4 0.2

0

NLP 101 NLM 102 NLM 202

200

400

600

800

1000

1200

1400

Temperature (°C)

Figure 5.111 The strengths of three commercially available Nicalon  fibers as a function of temperature. Reprinted, by permission, from A. Kelly, ed., Concise Encyclopedia of Composite Materials, p. 256, revised edition. Copyright  1994 by Elsevier Science Publishers, Ltd.

5.4.2.7 Matrix Cracking in Ceramic-Matrix Composites∗ . Silicon carbide fibers and whiskers also find common usage in ceramic-matrix composites (CMCs). However, unlike their use in matrices that undergo plastic deformation such as polymers and metals, in which a strong interfacial bond between the fiber and the matrix is desirable in order to facilitate load transfer, their use in brittle matrix composites such as CMCs is based upon the ability of the fiber to pull away from the matrix in a controlled manner in what is called fiber pull-out or graceful failure. The tensile modulus of a unidirectional continuous-fiber-reinforced ceramic-matrix composite is given to a good approximation by the rule of mixing [Eq. (5.119)]. The strength, however, is more complex than the rule of mixing in brittle-matrix composites. A successful continuous, unidirectional fiber-reinforced CMC will have fibers whose failure strain is significantly greater than that of the matrix and are not too strongly bonded to the matrix. The matrix thus fractures at a stress, σmu , that is lower than the ultimate failure stress of the composite, σcu . The typical stress–strain behavior for a unidirectional-reinforced CMC stressed in the axial fiber direction is shown in Figure 5.112. For a matrix material with a reasonable density of inherent flaws and zero fiber/matrix interfacial shear strength, τi , the curve would follow the path OACD, where the segment AC represents fiber pull-out as the matrix fails. In more realistic situations, the curve OAD is followed. Matrix cracking and the ultimate composite strength can be analyzed using a number of approaches, but we will utilize the fracture mechanics approach so that we can compare the results with those we developed in Sections 5.2.1.2 and 5.2.3.1 for

506

MECHANICS OF MATERIALS

B

STRESS

scu

A

smu

EC

0

D

C

EfVf

STRAIN

Figure 5.112 Idealized stress–strain curve for a tough ceramic-matrix composite. Reprinted, by permission, from R. W. Davidge and J. J. R. Davies, in Mechanical Testing of Engineering Ceramics at High Temperatures, B. F. Dyson, R. D. Lohr, and R. Morrell, eds., p. 251. Copyright  1989 by Elsevier Science Publishers, Ltd.

REGION 1

sa

REGION 2

sa REGION 3

C* C sa

sa

Figure 5.113 Microcrack parallel to fibers under applied stress, σa . Reprinted, by permission, from R. W. I. Davidge, and J. J. R. Davies, in Mechanical Testing of Engineering Ceramics at High Temperatures, B. F. Dyson, R. D. Lohr, and R. Morrell, eds., p. 253. Copyright  1989 by Elsevier Science Publishers, Ltd.

fracture in monolithic ceramics. Consider the unidirectional, fiber-reinforced CMC in Figure 5.113. A crack of length c has passed some way through the material, creating three distinct regions. Region 3 is remote from the crack, and is unaffected by it. At the mouth of the crack (Region 1), the crack is fully established and the crack faces have relaxed to a constant crack opening displacement. Near the crack tip (Region 2) there is a transition between the outer two regions. It has been shown that the stress

MECHANICS OF COMPOSITES

507

required to propagate the crack, σm , is given by 12τ γm σm = r



Vf2 Ef (1 − ν 2 )2 Vm Ec Em2

1/3

Ec 1 − ν2

(5.126)

where τ is the fiber–matrix interfacial frictional force, r is the fiber radius, γm is the surface energy of the matrix, ν is Poisson’s ratio of the matrix phase, Vf is the volume fraction of fibers, and Ec , Ef , and Em are the elastic moduli of the composite, fiber, and matrix, respectively. This situation is to be contrasted with that for the unreinforced material where the stress at the crack tip increases with an increase in crack length [Eq. (5.45)]. For these composite systems, the stress at the crack tip is independent of the crack length once the crack is longer than the equilibrium length, c. Thus, matrix cracking is impossible below the threshold stress, σm , no matter how large the preexisting defect. Composites with intrinsic flaws less than c will be expected to have a matrix cracking stress greater than σm , as indicated in Figure 5.114. The threshold crack size is usually several times larger than the average fiber spacing, which is of the order of the inherent flaw size. At stresses greater than σmu , the matrix will fracture into a series of parallel-sided blocks. The ultimate strength of the composite depends on the strength of the fibers σfu , such that σcu = σfu Vf (5.127)

MATRIX CRACKING STRESS (s /s ) mu a APPLIED STRESS

Note that at σcu , some of the load is still carried by the matrix in the unbroken regions and that the maximum stress on the fibers is restricted to the bridging regions between single matrix cracks. Thus, σfu may be greater than the value obtained for unrestrained fibers. The flexural and fracture properties of some common CMCs are given in Table 5.10, compared to their unreinforced matrix materials. Notice that the SiC fibers described

2.0

1.5 1.33

1.0

0

10

20

30

NORMALIZED CRACK LENGTH (C/Co)

Figure 5.114 Matrix cracking as a function of normalized crack length in CMCs. Reprinted, by permission, from R. W. Davidge and J. J. R. Davies, in Mechanical Testing of Engineering Ceramics at High Temperatures, B. F. Dyson, R. D. Lohr, and R. Morrell, eds., p. 254. Copyright  1989 by Elsevier Science Publishers, Ltd.

508

MECHANICS OF MATERIALS

Table 5.10 Comparison of Mechanical Properties of CMCs and the Equivalent Unreinforced Ceramic Materiala Unreinforced Al2 O3 30 vol% SiC whisker-Al2 O3 20 vol% SiC fibers (C coated)-Al2 O3 20 vol% SiC fibers (SiO2 coated)-Al2 O3 Unreinforced LAS glass ceramic 50 vol% 1-D SiC fiber LAS HPSN 30 vol% SiC-whisker HPSN Unreinforced RBSN 20 vol% 1-D SiC fiber RBSN

Bend Strength (MPa)

K1c (MPam1/2 )

385–600 650–800

4–5 6–9.5 8.3 7.2 2 17 3.4–8.2 10.5 1.5–2.8 54

200 600 400 450 200–400 650

a

HP, hot pressed; RB, reaction bonded; LAS, lithium aluminosilicate. Source: W. E. Lee, and W. M. Rainforth, Ceramic Microstructures. Copyright  1994 by William E. Lee and W. Mark Rainforth.

in the previous section for use in MMCs are also commonly used for CMCs; also notice that just as for MMCs, processing methods figure prominently in dictating the ultimate properties of the CMC. One ceramic-matrix material worthy of further description is a glass ceramic (cf. Section 1.2.5), such as lithium aluminosilicate (LAS). In these CMCs, the major crystalline phase is β-spudomene, and the useful temperature range of the composite is 10–1200◦ C. The SiC–LAS composite can be formed by either of two common techniques: filament winding or hot pressing. As with all glass-ceramics, controlled crystallization of the matrix phase is necessary to obtain the glass-ceramic structure (see Figure 3.17). Unidirectional reinforcement is unlikely to be used because the strength normal to the fibers is extremely low (20 MPa; see curve 1D 90◦ in Figure 5.115). For most practical applications at least two-dimensional reinforcement is used. Some temperature-dependent mechanical property data are available, as presented in Figure 5.116 for the variation of bend strength with temperature for the one-dimensional, two-dimensional, and unreinforced LAS glass ceramic. An interesting characteristic of the two-dimensional CMC is that there is a significant increase in bending strength around 1000◦ C due to oxidation, which increases the fiber–matrix interfacial shear strength. 5.4.3

Mechanical Properties of Laminate Composites

At the end of Section 5.4.2.5, the statement was made that most continuous, unidirectional fiber-reinforced composites are used to produce layers that are subsequently assembled to form laminate composites. In this section, we expound upon this statement by examining the mechanics of laminate composites, first through a generalized description of their mechanics, then with some specific stress–strain behavior. In the case of unidirectionally reinforced layers, or laminae (not to be confused with the similar-looking lamellae, cf. Section 1.3.6.1), a composite is formed by laminating the layers together. If all the layers were to have the same fiber orientation (see Figure 5.117a), the material would still be weak in the transverse direction. Therefore,

MECHANICS OF COMPOSITES

509

600

Tensile stress (MPa)

500

1D 0°

400 300 200 2D 0°/90° 100 2D 45°/45° 0

1D 90° 0.2

0

0.4

0.6

0.8

1.0

1.2

Strain (%)

Figure 5.115 Stress–strain diagrams for lithiumaluminosilicate glass ceramic reinforced with 50% SiC fibers in various orientations. From Ceramic Microstructures, by W. E. Lee and W. M. Rainforth, p. 103. Copyright  1994 by William E. Lee and W. Mark Rainforth, with kind permission of Kluwer Academic Publishers.

BEND STRENGTH (MPa)

1000

800 UNIDIRECTIONAL COMPOSITE

600

CROSS-PLIED (0°/90°) COMPOSITE

400 MONOLITHIC LAS

200

0

0

100

600

800

1000

1200

TEMPERATURE (°C)

Figure 5.116 Temperature dependence of bending strength for SiC-reinforced lithium aluminosilicate (LAS) CMC. Reprinted, by permission, from R. W. Davidge and J. J. R. Davies, in Mechanical Testing of Engineering Ceramics at High Temperatures, B. F. Dyson, R. D. Lohr, and R. Morrell, eds., p. 264. Copyright  1989 by Elsevier Science Publishers, Ltd.

angle-ply laminates are preferred for most applications. Typically, the laminae are stacked so that the fiber directions are symmetrical about the mid-plane of the composite, as in Figure 5.117b. A symmetrical stacking sequence ensures that in-plane stresses do not produce out-of-plane twisting or bending, as in Figure 5.118. The stacking sequence can be either 0◦ /90◦ , as in the bottom two plies of Figure 5.117b,

510

MECHANICS OF MATERIALS







90°



+45°



−45°



−45°



+45°



90°





UNIDIRECTIONAL (a)

CROSSPLIED QUASI−ISOTROPIC (b)

Figure 5.117 Schematic illustration of unidirectional (a) and crossplied (b) laminar composites. Reprinted, by permission, from M. Schwartz, Composite Materials Handbook, 2nd ed., p. 3.71. Copyright  1992 by McGraw-Hill, Inc.

Figure 5.118 Twisting, bending, and shear of an unsymmetrical laminate under direct stress. Reprinted, by permission, from P. C. Powell and A. J. I. Housz, Engineering with Polymers, p. 201. Copyright  1998 by Stanley Thorned Publishers.

termed cross-ply, −45◦ /+45◦ as in the third and fourth ply of Figure 5.117b, termed angle-ply, or a combination of the two. The layers can be of the same thickness, or different thickness, and multiple layers may be oriented in the same direction before changing the orientation of the next multiple layers—for example, three layers at 0◦ , three layers at +45◦ , three layers at −45◦ , and so on. The outer layers of the composite may be selected to provide a protective coating, as described previously.

MECHANICS OF COMPOSITES

2

511

y

q

1 x

Figure 5.119 Cartesian coordinate system relative to principle fiber directions in unidirectional fiber composite. Reprinted, by permission, from P. C. Powell and A. J. I. Housz, Engineering with Polymers, p. 215. Copyright  1998 by Stanley Thorned Publishers.

5.4.3.1 Stress–Strain Behavior for Unidirectional Laminae∗ . In Section 5.1.1.2, we developed a generalized view of elastic response (Hooke’s Law) that included all stresses and strains within the framework of a matrix notation. Equation (5.6) gave the stress matrix, and Eq. (5.8) gave the matrix of modulus components, Qij . Let us relate the generalized Hooke’s Law model to one lamina of continuous, unidirectional fiberreinforced matrix, as first illustrated in Figure 5.85. The directions 1, 2, and 3, defined in terms of the fiber orientation direction, can be related to the Cartesian coordinates x, y, and z as in Figure 5.119, through the introduction of the orientation angle, θ . We will retain the coordinates 1, 2, and 3 to relate relationships developed in this section to those previously presented for unidirectional composites. Equation (5.7) related the stress tensor to the strains through the modulus matrix, [Q], also known as the reduced stiffness matrix. Oftentimes, it is more convenient to relate the strains to the stresses through the inverse of the reduced stiffness matrix, [Q]−1 , called the compliance matrix, [S]. Recall from Section 5.3.1.3 that we introduced a quantity called the compliance, which was proportional to the inverse of the modulus. The matrix is the more generalized version of that compliance. The following relationship then holds between the strain and stress tensors:

ε = [S]σ

(5.128)

Let us first consider the case of an isotropic material, then simplify it for the case of an orthotropic material (same properties in the two directions orthogonal to the fiber axis—in this case, directions 2 and 3), such as a unidirectionally reinforced composite lamina. Equation (5.128) can be written in terms of the strain and stress components, which are coupled due to the anisotropy of the material. In order to describe the behavior in a manageable way, it is customary to introduce a reduced set of nomenclature. Direct stresses and strains have two subscripts—for example, σ11 , ε22 , τ12 , and γ21 , depending on whether the stresses and strains are tensile (σ and ε) or shear (τ and γ ) in nature. The moduli should therefore also have two subscripts: E11 , E22 , and G12 , and so on. By convention, engineers use a contracted form of notation, where possible, so that repeated subscripts are reduced to just one: σ11 becomes σ1 , E11 becomes E1 , but G12 stays the same. The convention is further extended for stresses and strains, such that distinctions between tensile and shear stresses and strains are

512

MECHANICS OF MATERIALS

dropped. So, the coupling between stresses and strains in an anisotropic material can now be expressed in the generalized Hooke’s Law form:   ε1 S11  ε2   S21     ε3   S31  =  γ31   S41 γ  S 32 51 γ12 S61 

S12 S22 S32 S42 S52 S62

S13 S23 S33 S43 S53 S63

S14 S24 S34 S44 S54 S64

S15 S25 S35 S45 S55 S65

  σ1 S16 S26   σ2    S36   σ3    S46   τ31    S56 τ32  S66 τ12

where [S] is the symmetric compliance matrix. For an anisotropic lamina under in-plane loads, Eq. (5.129) reduces to      ε1 S11 S12 S16 σ1  ε2  =  S21 S22 S26   σ2  γ12 S61 S62 S66 τ12

(5.129)

(5.130)

and for an orthotropic lamina, such as in Figure 5.119, under in-plane stresses in the principal directions, Eq. (5.130) further reduces to      ε1 S11 S12 0 σ1  ε2  =  S21 S22 0   σ2  (5.131) γ12 0 0 S66 τ12

where the subscripts 1 and 2 refer to the fiber and transverse directions, respectively. The elastic compliance constants, Sij (often called the engineering constants) can be expressed in terms of unidirectional lamina properties: S11 = 1/E1 ; S12 = S21 = −ν12 /E1 = −ν21 /E2 ; S22 = 1/E2 , and S66 = 1/G12 , where ν is Poisson’s ratio. Equation (5.131) can be inverted to give the reduced stress response to the inplane stress:      Q11 Q12 σ1 0 ε1  σ2  =  Q21 Q22 0   ε2  (5.132) τ12 0 0 Q66 γ12

where again [Q] = [S]−1 , and relationships for each Qij can be similarly expressed in terms of the lamina moduli; for example, Q66 = G12 . The next level of complexity in the treatment is to orient the applied stress at an angle, θ , to the lamina fiber axis, as illustrated in Figure 5.119. A transformation matrix, [T ], must be introduced to relate the principal stresses, σ1 , σ2 , and τ12 , to the stresses in the new x –y coordinate system, σx , σy , and τxy ; and the inverse transformation matrix, [T ]−1 , is used to convert the corresponding strains. The entire development will not be presented here. The results of this analysis are that the tensile moduli of the composite along the x and y axes, Ex , and Ey , which are parallel and transverse to the applied load, respectively, as well as the shear modulus, Gxy , can be related to the lamina tensile modulus along the fiber axis, E1 , the transverse tensile modulus, E2 , the lamina shear modulus, G12 , Poisson’s ratio, ν12 , and the angle of lamina orientation relative to the applied load, θ , as follows:   sin4 θ cos4 θ 2ν12 1 1 = + − (5.133) sin2 θ cos2 θ + Ex E1 G12 E1 E2

513

MECHANICS OF COMPOSITES

sin4 θ 1 = + Ey E1



2ν12 1 − G12 E1



sin2 θ cos2 θ +

cos4 θ E2

(5.134)

and   1 sin4 θ + cos4 θ 2 4ν12 1 2 sin2 θ cos2 θ + =2 + + − Gxy E1 E2 E1 G12 G12

(5.135)

Poisson’s ratio for the off-axis loaded lamina, νxy , can also be derived. The relative tensile modulus, Ex /E1 , shear modulus, Gxy /G12 , and Poisson’s ratio, νxy , are plotted as a function of the angle of rotation, θ , for a glass-fiber-reinforced epoxy lamina and a graphite-fiber-reinforced epoxy lamina in Figures 5.120a and 5.120b, respectively. The final level of complexity is to arrange F layers of unidirectional laminae to form the laminate composite, as shown in Figure 5.121. The lower surface of the f th lamina is assigned the coordinates hf relative to the midplane, so the thickness of the f th lamina is hf − hf −1 . The mathematics of stress analysis are an extension of those used in the foregoing description, and they include force and moment analysis on the bending of the laminate. The complete development will not be presented here, but only outlined. The interested reader is referred to two excellent descriptions of the results [9,10]. The laminate forces and moments are obtained by integrating the stresses of each lamina through the total laminate thickness to obtain a force vector, N, and a moment vector, M, the components of which can be related to the strains, ε0 and γ 0 , and curvatures, κ, of the laminae, where the superscript indicates that the strains are measured

Gxy /G12 1.5

Gxy /G12

nxy

0.3

0.6 0.4

0.2

Ex El

0.1

0.2 0

30 60 q (deg) (a)

0 90

1.0

0.3 nxy 0.2

0.5

Ex El

0

nxy

0.8

Ex /El and Gxy /G12

1.0

nxy

Ex /El and Gxy /G12

1.2

0.1

30 60 q (deg)

90

(b)

Figure 5.120 Elastic constants for unidirectional (a) glass fiber reinforced epoxy and (b) graphite fiber-reinforced epoxy laminae. Reprinted, by permission, from P. C. Powell and A. J. I. Housz, Engineering with Polymers, pp. 222, 223. Copyright  1998 by Stanley Thorned Publishers.

514

MECHANICS OF MATERIALS

1 2

h2

h0

h1 hf

hF −1 hF

f F

Figure 5.121 Thickness coordinates in a laminate composite. Reprinted, by permission, from P. C. Powell and A. J. I. Housz, Engineering with Polymers, p. 231. Copyright  1998 by Stanley Thorne Publishers.

relative to the midplane of the laminate:  0    εx κx Nx  Ny  = [A]  εy0  + [B]  κy  Nxy κxy γxy0  and

 0     εx Mx κx  My  = [B]  εy0  + [D]  κy  Mxy κxy γxy0 

(5.136)

(5.137)

where the arrays [A], [B], and [D] are given by Aij =

F

Bij =

1 (Qij )f (h2f − h2f −1 ) 2 f =1

Dij =

1 (Q )f (h3f − h3f −1 ) 3 f =1 ij

f =1

(Qij )f (hf − hf −1 )

(5.138)

F

and

(5.139)

F

(5.140)

Here, the array [Q]f is the transformed reduced stiffness matrix, and each Qij term can be related to the Qij terms of the reduced stiffness matrix, [Q], through the angle, θ . 5.4.3.2 Strength of Laminate Composites. The strength of angle-ply laminates can be calculated using the same type of elastic analysis as for modulus in the previous section. The strains produced in the laminate by a given set of applied stresses are first calculated, using the computed laminate moduli. Stresses corresponding to these strains are then calculated for each layer of the laminate. These stresses are then expressed in terms of stresses parallel and normal to the fibers, and the combination of stresses

MECHANICS OF BIOLOGICS

1000

q° 90 80 70 60 50 40 30 20 10 0

515

0

Tensile strength (MPa)

0.2

0.4 500 0.6

0.8

0

1.0

R

Figure 5.122 Strength chart for a (0, ±θ ) unidirectional carbon-fiber-reinforced epoxy laminate composite. Reprinted, by permission, from N. G. McCrum, C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, 2nd ed., p. 408. Copyright  1997 by Oxford University Press.

is compared with strength criteria for the unidirectional material. Results for a threedirection carbon-fiber-reinforced epoxy composite (0, ±θ ) under tensile stress in the θ = 0◦ direction are shown in Figure 5.122. The composite contains 60 vol% fibers. Here, R is the fraction of plies which are angled at ±θ , and 1-R is the fraction at θ = 0. In cross-ply laminates, the stress–strain behavior is slightly nonlinear, as illustrated in Figure 5.123. The stress–strain behavior of a unidirectional lamina along the fiber axis is shown in the top curve, while the stress–strain behavior for transverse loading is illustrated in the bottom curve. The stress–strain curve of the cross-ply composite, in the middle, exhibits a knee, indicated by strength σk , which corresponds to the rupture of the fibers in the 90◦ ply. The 0◦ ply then bears the load, until it too ruptures at a composite fracture strength of σf . One additional problem area for fiber-reinforced thermosetting laminates is interlaminar splitting, which occurs particularly easily when the void content is high. Measurements of interlaminar shear strength (ISS) are made by subjecting short bar specimens to three-point bending. Deformation of laminated composites can also occur because of changes in temperature and the absorption of moisture. This is known as the hygrothermal effect, and it causes the polymer matrix to undergo both dimensional and property changes with temperature and humidity.

5.5

MECHANICS OF BIOLOGICS

Most of the terms, concepts, and equations for studying the mechanical properties of biological materials have already been developed in the previous sections of this

516

MECHANICS OF MATERIALS

s [0]

[0/90]S

sF ES

sk

KNEE

E [90] eTt

eLt

e

Figure 5.123 Schematic tensile stress–strain diagram for a symmetric (0, ±90◦ ) cross-plied laminate tested at 2 = 0◦ . Reprinted, by permission, from P. K. Mallick, Fiber-Reinforced Composites, p. 187. Copyright  1988 by Marcel Dekker, Inc.

chapter. What remains is to relate the unique structural characteristics of biologics to their mechanical properties. As in Chapter 1, we will divide the biologics into soft and hard classifications, since their structure and mechanical responses are uniquely different. These differences are summarized in Figure 5.124, in which the stress–strain curves of some soft and hard biologics are compared. The types of materials used as substitutes for these two classifications are also different. In each case, we will begin by examining the structure and mechanical properties of true biological materials, then briefly describe materials that can be used to replace them. 5.5.1

Mechanical Properties of Soft Biologics

Because most biological structures are composed of α helices, β sheets (cf. Section 1.5.1.1), and collagen triple helixes, it is important to understand the mechanical properties of these basic structures. For example, the mechanical properties of tendon collagen fibers are believed to reflect the mechanical properties of the collagen triple helix which serves as the building block for the collagen protein fiber (see Figure 5.125, cf. Figure 1.93). Ultimate tensile strength values as high as 40 MPa, strains up to 10%, and high strain moduli in excess of 500 MPa have been reported for isolated collagen fibers with diameters from 50 to 100 µm. These properties are related to the stiffness of the triple helix, which has been estimated to have an elastic modulus as high as 4 GPa. Although the mechanical properties of connective tissues such as tendons and ligaments are dependent upon the properties of the collagen fibrils, their diameters, and the content of the collagen in the tissue, they are also dependent upon several other factors, including the orientation of the collagen fibers and the extent of tissue hydration. The latter is believed to occur due to the interaction of water with the hydrophobic regions of the collagen protein.

MECHANICS OF BIOLOGICS

517

Relative Stress

Bone

Tendon

Skin

0.2

0.4 Strain

0.6

Figure 5.124 Schematic comparison of stress–strain curves for some soft and hard biological materials. Reprinted, by permission, from F. H. Silver and D. L. Christiansen, Biomaterials Science and Biocompatibility, p. 18. Copyright  1999 by Springer-Verlag.

In tissues such as tendons that function to transmit loads along a single direction, the collagen fibers are oriented in only one direction. Ultimate tensile strengths as high as 120 MPa, along with strains of about 10%, have been reported for tendons. For ligaments, these values are 147 MPa and 70%, respectively (see Table 5.11). Other tissues, such as skin, bear loads within a plane and need to have collagen fibers oriented in the plane. In the next section, we look more closely at the structure–mechanical property relationships in skin. 5.5.1.1 Skin. Tissue such as tendons and ligaments are termed oriented networks because the collagen fibers are oriented in a preferred direction. Other collagencontaining tissues such as skin are termed orientable. Because skin normally bears loads within the plane of the skin, the collagen fibers become oriented with the direction of loading and reorient during mechanical deformation. Thus, skin is able to bear loads in any direction within its plane. For this reason, the collagen network in skin must be loose enough to be able to rearrange. Other types of soft tissue such as aorta and cartilage contain some orientable collagen networks. The stress–strain curve for skin is shown in Figure 5.126. For strains up to about 30%, the collagen network offers little resistance to deformation, which is similar to

518

MECHANICS OF MATERIALS

Lateral aggregation 4 nm Formation of microfibrils

1.5 nm Collagen molecule End-to-end aggregation

Collagen fiber

Figure 5.125 Structure of collagen fibers based on collagen molecule, microfibrils and fibrillar aggregation. Reprinted, by permission, from I. V. Yannas, in Biomaterials Science, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, eds., p. 86. Copyright  1996 by Academic Press.

Table 5.11 Selected Mechanical Properties of Some Soft Biologics

Tissue Arterial wall Tendons/ligaments Skin Hyaline cartilage

Maximum Strength (MPa)

Maximum Strain (%)

Elastic Modulus (MPa)

0.24–1.72 50–150 2.5–15 1–18

40–53 5–50 50–200 10–120

1.0 100–2500 10–40 0.4–19

Source: F. H. Silver and D. L. Christiansen, Biomaterials Science and Biocompatibility. Copyright  1999 by Springer-Verlag.

tendon in this region. Collagen fibers are initially wavy and become straight upon loading. Between 30% and 60% strains, the collagen fibers become aligned and offer resistance to further extension. Over 60% elongation, the fibers begin to yield and fail. At high strains, the modulus is about 22.4 MPa, and the ultimate tensile strength is 14 MPa, based on the true cross-sectional area. The elastic fraction of skin is defined as the ratio of the total force divided by the force at equilibrium and is a measure of the viscoelastic nature of a material. For completely elastic materials, the elastic fraction is 1.0, whereas for viscoelastic materials, the value lies between 0 and 1.0. In skin, the elastic fraction increases from about 0.50 to 0.73 with increasing strain, as illustrated in Figure 5.127. This is in contrast to other types of collagen-containing soft tissue such as aorta.

MECHANICS OF BIOLOGICS

25

519

Linear Region 50% Strain rate

20

Stress (kg/cm 2 )

10% Strain rate 15

10

Low Modulus Region

5

0

10

20

30

40

50 60 Strain (%)

70

80

90

100

Figure 5.126 Stress–strain curves for wet back skin at various strain rates. Reprinted, by permission, from F. H. Silver and D. L. Christiansen, Biomaterials Science and Biocompatibility, p. 203. Copyright  1999 by Springer-Verlag.

Elastic Fraction

0.9

0.8

0.7

0.6

0.5

0.4

0

10

20

30

40

50 Strain (%)

60

70

80

90

100

Figure 5.127 Effect of strain on elastic fraction of aorta in the circumferential (solid triangles) and axial (open squares) directions, as well as on skin (open circles). Reprinted, by permission, from F. H. Silver and D. L. Christiansen, Biomaterials Science and Biocompatibility, p. 202. Copyright  1999 by Springer-Verlag.

520

MECHANICS OF MATERIALS

The mechanical behavior of skin is HISTORICAL HIGHLIGHT time-dependent and strain-rate-dependent, The directional dependence of the and for this reason the elastic fraction is mechanical properties of skin was first less than 1.0. The time dependence of observed by Langer in 1861. C. R. von the mechanical properties and the strain E. Langer (1819–1887), an Austrian dependence of the elastic fraction imply an anatomist, pierced small circular holes interesting structural transition that occurs in the skin of cadavers and observed in skin with strain. This behavior is that these holes became elliptical with attributable to the collagen network. At time. By joining the long axes of the low strains, the elastic fraction is low and ellipses, he produced a series of lines that appears to involve the viscous realignment subsequently were named after him. It of the collagen fibers with the applied is well known to surgeons that incisions made along the direction of Langer’s stress. At higher strains, the elastic fraction lines close and heal more rapidly. approaches 0.75, which is near the value for tendon material. In fact, several types of Source: Silver, Oxford Concise Medical Dictionary connective tissue exhibit an elastic fraction approaching 0.75 at high strains. This suggests that the collagen network limits the high strain deformation in all types of connective tissues. Although the proteins in skin are also composed of about 5% elastic fibers, they do not appear to affect the mechanical properties of the tissue. The elastic fibers are believed to contribute to the recoil of the skin, which gives it the ability to be wrinklefree when external loads are removed. As humans age, the elastic fiber network of the skin is lost, and wrinkles begin to appear. The mechanical role of the elastic fibers is very different in vascular tissue, however. Fatigue testing of skin in uniaxial tension shifts the stress–strain curve to the right within the linear region, and the elastic fraction approaches 0.85. Therefore, mechanical tests must be carried out carefully to eliminate any preconditioning effects. Further complicating testing is the fact that the mechanical properties of skin depend on the location from which the specimen is obtained. The contribution of the epithelial layer to the mechanical properties depends on its thickness. For most regions, the stratum corneum is thin and does not contribute significantly to the mechanical properties of skin; however, in places where it is thick, such as the palms of the hand, it needs to be considered. 5.5.1.2 Sutures. A suture is a biomedical product used to hold tissue together following separation by surgery or trauma. Sutures are broadly categorized according to the type of material (natural, synthetic) from which they are made, the permanence of the material (absorbable or nonabsorbable), and the construction process (braided, monofilament). As shown in Table 5.12, the most popular natural materials used for sutures are silk and catgut (animal intestine). The synthetic materials are exclusively polymeric, except for fine-sized stainless steel sutures. Approximately half of today’s sutures are nonabsorbable and remain indefinitely intact when placed in the body. Polymers such as polypropylene, nylon, poly(ethylene terephthalate), and polyethylene, as well as their copolymers, are used as sutures. Absorbable sutures were commercially introduced with poly(glycolic acid) (PGA), and they were followed by copolymers of glycolide and lactide. A PGA suture will lose strength over

MECHANICS OF BIOLOGICS

521

Table 5.12 Representative Mechanical Properties of Some Commercial Sutures

Suture Type Natural materials Catgut Silk Synthetic absorbable Poly(glycolic acid) Poly(glycolide-co-lactide) Poly(p-dioxanone) Poly(glycolide-co-trimethylene carbonate) Synthetic nonabsorbable Poly(butylene terephthalate) Poly(ethylene terephthalate) Poly[p(tetramethylene ether) terephthalate-co-tetramethylene terephthalate] Polypropylene Nylon 66 Steel

St. Pull (MPa)

Kt. Pull (MPa)

Elongation to Break (%)

Subjective Flexibility

370 470

160 265

25 21

Stiff Very supple

840 740 505 575

480 350 290 380

22 22 34 32

Supple Supple Moderately stiff Moderately stiff

520 735 515

340 345 330

20 25 34

Supple Supple Supple

435 585 660

300 315 565

43 41 45

Stiff Stiff Rigid

Source: D. Goupil, in Biomaterials Science, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, ed., p. 356. Copyright  1996 by Academic Press.

a 28-day period. More recently, novel absorbable polymers of polydioxanone and poly(glycolide-co-trimethylene carbonate) have been developed for surgical use. Regardless of whether a suture is natural or synthetic, or if it is absorbable or permanent, it must meet the strength requirements necessary to close a wound under a given clinical circumstance. Suture strength is generally measured and reported in both straight pull and knotted pull tests. The former is essentially a traditional tensile test; the second is obtained after the suture has been knotted. As illustrated in Table 5.12, knotted pull strengths are approximately one-half to two-thirds of the straight pull strength values. There are a number of other characteristics a suture must have in addition to strength, including knot security, strength retention, resistance to infection, flexibility, and ease of tying. 5.5.1.3 Artificial Soft Biologics. In addition to sutures, polymers are used for a number of biomedical applications, as illustrated in Figure 5.128. Polymers used for hard structural applications such as dentures and bones are presented in this figure, but will be described in the next section. In this section, we will concentrate on polymers for soft biological material applications and will limit the description to mechanical properties as much as possible. Artificial biologics, whether soft or hard, can be categorized as either temporary (short term) or permanent (long term) in their intended application. Most, but certainly not all, polymers for biomedical applications are of the short-term type and include sutures, drug delivery devices, temporary vascular grafts and stents, tissue scaffolds,

522

MECHANICS OF MATERIALS

Ear & ear parts Facial prosthesis Esophagus Lung, kidney & liver parts Heart pacemaker Gastrointestinal segments

Dentures Tracheal tubes Heart & heart components Biodegradable sutures Finger joints

Blood vessels Knee joints Bones & joints

Ear & ear parts: acrylic, polyethylene, silicone, poly(vinyl chloride) (PVC) Dentures: acrylic, ultrahigh molecular weight polyethylene (UHMWPE), epoxy Facial prosthesis: acrylic, PVC, polyurethane (PUR) Tracheal tubes: acrylic, silicone, nylon Heart & heart components: polyester, silicone, PVC Heart pacemaker: polyethylene, acetal Lung, kidney & liver parts: polyester, polyaldehyde, PVC Esophagus segments: polyethylene, polypropylene (PP), PVC Blood vessels: PVC, polyester Biodegradable sutures: PUR Gastrointestinal segments: silicones, PVC, nylon Finger joints: silicone, UHMWPE Bones & joints: acrylic, nylon, silicone, PUR, PP, UHMWPE Knee joints: polyethylene

Figure 5.128 Common applications of polymers as soft biologics. Reprinted, by permission, from S. A. Visser, R. W. Hergenrother, and S. L. Cooper, in Biomaterials Science, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, eds., p. 57. Copyright  1996 by Academic Press.

and orthopedic fixation devices. Short-term devices can either be removable or degradable. We will concentrate for the moment on biodegradable polymers and will make no distinction between often-confused terms such as biodegradable, bioresorbable, bioerodable, and bioabsorbable. In all cases, we describe a polymer that is converted under physiological conditions into water-soluble materials. Some of the common biodegradable polymers used in short-term biomedical applications, and their corresponding mechanical properties, are listed in Table 5.13. Polyhydroxybutyrate (PHB), polyhydroxyvalerate (PHV), and their copolymers are derived from microorganisms and are used in such applications as controlled drug release, sutures, and artificial skin. Polycaprolactone is a semicrystalline polymer that is used in some applications as a degradable staple. Polyanhydrides are very reactive and hydrolytically unstable, which is both an advantage and a limitation to their use. A wide variety of drugs and proteins, including insulin, bovine growth factors, heparin, cortisone, and enzymes, have been incorporated into polyanhydride matrices. Poly(ortho esters) have also been used for controlled drug delivery. Poly(amino acids) are of

MECHANICS OF BIOLOGICS

523

Table 5.13 Selected Mechanical Properties of Some Biodegradable Polymers

Polymer Poly(glycolic acid) (MW: 50,000) Poly(lactic acids) L-PLA (MW: 50,000) L-PLA (MW: 100,000) L-PLA (MW: 300,000) D, L-PLA (MW: 20,000) D, L-PLA (MW: 107,000) D, L-PLA (MW: 550,000) Poly(β-hydroxybutyrate) (MW: 422,000) Poly(ε-caprolactone) (MW: 44,000) Polyanhydrides Poly(SA-HDA anhydride) (MW: 142,000) Poly(ortho esters) DETOSU: t-CDM: 1,6-HD (MW: 99,700) Polyiminocarbonates Poly(BPA iminocarbonate) (MW: 105,000) Poly(DTH iminocarbonate) (MW: 103,000)

Elongation Glass Melting Tensile Tensile Flexural Transition Temperature Strength Modulus Modulus Yield Break (◦ C) (◦ C) (MPa) (MPa) (MPa) (%) (%) 35

210

n/a

n/a

n/a

n/a

n/a

54 58 59 50 51 53 1

170 159 178 — — — 171

28 50 48 n/a 29 35 36

1200 2700 3000 n/a 1900 2400 2500

1400 3000 3250 n/a 1950 2350 2850

3.7 2.6 1.8 n/a 4.0 3.5 2.2

6.0 3.3 2.0 n/a 6.0 5.0 2.5

−62

57

16

400

500

7.0

80

n/a

49

4

45

n/a

55



20

820

950

69



50

2150

2400

3.5

4.0

55



40

1630

n/a

3.5

7.0

14

85

4.1 220

Source: J. Kohn, and R. Langer, in Biomaterials Science, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, ed., p. 64. Copyright  1996 by Academic Press.

interest due to the presence of amino acid groups that can potentially be modified, but have found few biomedical applications because they are highly insoluble and difficult to process. Many of these biodegradable polymers, as well as nonbiodegradable fibers, are fabricated into fabrics for a variety of biomedical applications. An example of how these materials can be assembled into a useful artificial soft biologic is artificial skin. The use of artificial skin allows for prompt closure of the largest of burn wounds, thereby limiting wound contracture and providing permanent coverage with a readily available, easily stored skin replacement material. Key characteristics of artificial skin are biocompatibility and resistance to infection. Several types of synthetic skin grafts exist, but are all essentially bilaminate membranes, as shown in Figure 5.129. A small pore diameter is required in the outer layer to prevent the wound from drying out. However, a large pore diameter is required in the lower layer to promote the ingrowth of fibroblasts and long-term adhesion to the wound. One such bilaminate artificial skin consists of an inner layer of collagenchondroitin-6-sulfate and an outer layer of silastic. The collagen binds to the wound bed and is invaded by neovasculature and fibroblasts. The collagen chondroitin 6-sulfate very significantly inhibits wound contracture, and with time the collagen is resorbed and is completely replaced by remodeled dermis. The pore size of the collagen layer is typically 50 µm. The silicone layer is 0.1 mm thick and provides a bacterial barrier in addition to the required water flux rate of 1 to 10 mg/cm2 ·hr. The silicone layer

524

MECHANICS OF MATERIALS

Body Moisture

Silicone Layer

Epithelial cells

Biodegradable protein layer Fibroblasts, Endothelial Cells

Figure 5.129 Schematic diagram of bilaminar artificial skin. Reprinted, by permission, from J. B. Kane, R. G. Tompkins, M. L. Yarmush, and J. F. Burke, in Biomaterials Science, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, eds., p. 367. Copyright  1996 by Academic Press.

also provides the mechanical rigidity needed to suture the graft in place, thus preventing movement of the material during wound healing. The collagen layer adheres to the wound rapidly, as evidenced by the peel strength of the graft, which is 9 N/m at 24 hours and 45 N/m at 10 days. Histological cross sections of wounds closed with this artificial skin show complete replacement of the bovine collagen at seven weeks with remodeled human dermis. 5.5.2

Mechanical Properties of Hard Biologics

In contrast to soft biologics, whose mechanical properties primarily depend upon the orientation of collagen fibers, the mechanical properties of mineralized tissues, or hard biologics, are more complicated. Factors such as density, mineral content, fat content, water content, and sample preservation and preparation play important roles in mechanical property determination. Specimen orientation also plays a key role, since most hard biologics such as bone are composite structures. For the most part, we will concentrate on the average properties of these materials and will relate these values to those of important, man-made replacement materials. 5.5.2.1 Bone. The structure of bone was described in Section 1.5.2. Recall that bone is a composite material, composed primarily of a calcium phosphate form called hydroxyapatite (HA). The major support bones consist of an outer load-bearing shell of cortical (or compact) bone with a medullary cavity containing cancellous (or spongy) bone toward the ends. The stress–strain curves for cortical bones at various strain rates are shown in Figure 5.130. The mechanical behavior is as expected from a composite of linear elastic ceramic reinforcement (HA) and a compliant, ductile polymer matrix (collagen). In fact, the tensile modulus values for bone can be modeled to within a factor of two by a rule-of-mixtures calculation on the basis of a 0.5 volume fraction HA-reinforced

MECHANICS OF BIOLOGICS

525

de/dt = 1500/sec

Stress (MPa)

300

de/dt = 300/sec

de/dt = 1/sec

200

de/dt = 1.0/sec de/dt = 0.01/sec

de/dt = 0.001/sec 100

0

0.004

0.008 Strain

0.012

0.016

Figure 5.130 Stress–strain curves at various strain rates for cortical bone. Reprinted, by permission, from F. H. Silver and D. L. Christiansen, Biomaterials Science and Biocompatibility, p. 208. Copyright  1999 by Springer-Verlag.

collagen composite. The ultimate tensile strength for cortical bone varies from 100 to 300 MPa, the modulus varies up to 30 GPa, and the strain at failure is typically only 1% or 2%. Bone has either isotropic symmetry, requiring five stiffness coefficients (see Section 5.4.3.1 on generalized Hooke’s Law), or orthotropic symmetry, which requires nine such coefficients. These coefficients are particularly important for the modeling of bone resorption and implant performance. Figure 5.130 shows that at higher strain rates, the modulus increases and the bone behaves in a more brittle manner. Values for the critical stress intensity factor, K1c , of 2–12 MN/m3/2 and critical strain energy of 500–6000 J/m2 have been determined. Conversely, bone is much more viscoelastic in its response at very low strain rates. Recent experiments have shown that bone is a complex viscoelastic material; that is, time–temperature superposition cannot be used to obtain its properties. However, the nonlinearity in its mechanical behavior tends to occur outside the range of physiological interest, such that linear models are often useful for creep and stress relaxation behavior. Mineralization tends to increase modulus and ultimate tensile strength, as well as reducing the strain at failure. Mineralization does not affect the viscoelastic behavior of bone, however, and the increasing elasticity with decreasing strain rate observed in Figure 5.130 is followed, even with increased mineralization. The mechanical properties of cancellous bone are dependent upon the bone density and porosity, and the strength and modulus are therefore much lower than those for cortical bone. The axial and compressive strength are proportional to the square of the bone density, and moduli can range from 1 to 3 GPa.

526

MECHANICS OF MATERIALS

Table 5.14 Selected Mechanical Properties of Human Dentin and Enamel Dentin Property Density, g/cm3 Elastic modulus, GPa Tensile strength, MPa Compressive strength, MPa Shear strength, MPa Knoop hardness, kg/mm2 Poisson’s ratio

Mineralized

Demineralized

Enamel

2.14–2.18 14.7 105.5 297 138 68 0.31

— 0.26 29.6 — — — —

2.95–2.97 84–130 10.3 384 90.2 355–431 0.33

Source: Biomaterials Properties Database, W. J. O’Brien, director, www.lib.umich.edu/dentlib/Dental tables/toc.html

5.5.2.2 Teeth. The structure of a human tooth was given in Figure 1.91. Recall that the two most prominent components of teeth are enamel and dentin and that the primary components of each are hydroxyapatite (HA) and HA/collagen, respectively. The mechanical properties of dentin and enamel—and, where available, their demineralized forms—are presented in Table 5.14. Notice the severe decrease in modulus and tensile strength as dentin loses its mineral content. A knowledge and understanding of these properties are important for developing suitably compatible dental amalgams, adhesives, and prosthetics. 5.5.2.3 Artificial Hard Biologics. A summary of polymeric materials used in biomedical applications was given in Figure 5.128. Recall that many of these application were for replacement of soft tissues, such as skin and tendons. In this section, we concentrate not only on those polymeric materials used for hard biologics, but also on all classes of materials that find use in the bonding, replacing, and growth-scaffolding of hard biologics. A schematic summary of some of these applications that involve metallic, ceramic, polymeric and composite components is given in Figure 5.131. We will approach each of these materials classes in order and will describe only a few examples of the biomedical applications for each. In most cases, these applications will be as structural replacements, and in all cases we will concentrate on the specific mechanical properties that make each material suitable for the particular application. Topics such as biocompatibility, longevity, and infection resistance are beyond the scope of this text. Metals were some of the first materials to be used for the replacement of hard biologics. This was a reasonable selection, since the principal function of the long bones of the lower body is to act as load-bearing members. The hip prosthesis has been the most active area of joint replacement research. Both stainless steel, such as 316L, and Co–Cr alloys became the early materials of choice, since their stiffness, rigidity, and strength considerably exceed those of natural bone (see Table 5.15). Titanium and the so-called aviation alloy Ti–6Al–4V are now also used for the femoral portion of hip prostheses. In addition to strength and modulus, corrosion resistance and fatigue life are of prime importance. Also, the metals must be nontoxic, mutagenic, or carcinogenic within the body, either of themselves or due to the release of chemical components as a result of interactions with various body fluids. In fact, these are requirements for

MECHANICS OF BIOLOGICS

METALS Cranial Plates Ti, Ti Alloys 316L Stainless Steel Ta Orbit Reconstruction Co-Cr Mesh Ti Mesh Maxillofacial Reconstruction Ti Ti Mesh Ti-Al Valloy Dental implants 316L Stainless Steel Co-Cr-Mo Alloys Ti,Ti Alloys Bone Fracture Fixation 316L Stainless Steel Co-Cr-Mo Alloys Ti,Ti Alloys Heart Pacemaker Can: 316L Stainless Steel Ti Electrodes: Pt, Pt-lr Harrington Rod (spinal manipulation) Co-Cr-Mo Alloy 316L Stainless Steel Prosthetic Joints (hip, knee, shoulder, elbow, wrist) 316L Stainless Steel Co-Cr-Mo Alloys Ti Ti-Al-V-Alloys Harrington Rods Co-Cr-Mo Alloys

527

BIOCERMICS Cranial Repair Bioactive Glasses Keratoprostheses (Eye Lens) Al2O3 Otolaryngological implants Al2O3 HA Bioactive Glasses Bioactive Glass-Ceramics Maxillofacial Reconstruction Al2O3 HA HA-PLA Composite Bioactive Glasses Dental Implants Al2O3 HA Bioactive Glasses Aiveolar Ridge Augmentations Al2O3 HA; TCP HA-Autogenous Bone Composite HA-PLA Composite Bioactive Glasses Periodontal Pocket Obliteration HA HA-PLA Composite TCP Calcium and Phosphate Salts Bioactive Glasses Percutaneous Acces Devices Bioactive Glass-Ceramics Bioactive Glasses HA Pyrolytic Carbon Coating Artificial Heart Valves Pyrolytic Carbon Coatings Spinal Surgery Bioactive Glass-Ceramic HA Iliac Crest Repair Bioactive Glass-Ceramic Bone Space Fillers TCP Calcium and Phosphate Salts Bioactive Glass Granules Bioactive Glass-Ceramic Granules Orthopedic Load-Bearing Applications Al2O3 Stabilized Zirconia PE-HA Composite Coatings for Chemical Bonding (Orthopedic, Dental, and Maxillofacial Prosthetics) HA Bioactive Glasses Bioactive Glass-Ceramics Orthopedic Fixation Devices PLA-Carbon Fibers PIA-Calcium Phosphate-Based Glass Fibers Artificial Tendon and Ligament PLA-Carbon-Fiber Composite

Figure 5.131 Clinical uses of inorganic biomaterials. Reprinted, by permission, from L. Hench, in Materials Chemistry, L. V. Interrante, L. A. Casper, and A. B. Ellis, eds., p. 525. Copyright  1995 by American Chemical Society.

all classes of materials in these types of biomedical applications. In most prosthetic applications, the metal or alloy is coated with, and held in place by, polymers. Polymers have found two specific applications in orthopedics. First, they are used for one of the articulating surface components in joint prostheses. Thus, they must have a low coefficient of friction and low wear rate when they are in contact with the opposing surface, which is usually made of metal or ceramic. Initially, poly(tetrafluoroethylene), PTFE, was used in this type of application, but its accelerated creep rate and poor stress corrosion caused it to fail in vivo. Ultimately, this material was replaced with

528

MECHANICS OF MATERIALS

Table 5.15 Comparison of Some Mechanical Properties of Current Implant Materials with Those of Cortical Bone

Materials

Young’s Modulus (GPa)

Alumina Cobalt–chromium alloys Austenitic stainless steel Ti–6wt% Al–4wt% V Cortical bone PMMA bone cement Polyethylene

365 230 200 106 7–30 3.5 1

Ultimate Tensile Strength (MPa)

Critical Stress Intensity Factor K1C (MN m−3/2 )

6–55 900–1540 540–1000 900 50–150 70 30

∼3 ∼100 ∼100 ∼80 2–12 1.5

Critical Strain Energy Release Rate, G1C (J m−2 ) ∼40 ∼50,000 ∼50,000 ∼10,000 ∼600–5000 ∼400 ∼8000

Source: Concise Encyclopedia of Composite Materials, A. Kelly, ed., revised edition. Copyright  1994 by Elsevier Science Ltd.

ultrahigh-molecular-weight polyethylene (UHMWPE). The second-place polymers find biomedical application is as a structural interface between the implant and bone tissue. In this case, the appropriate mechanical properties of the polymer are of major importance. The first type of use was with PMMA as a grouting material to fix both the stem of the femoral component and the acetabular component in place and to distribute the loads more uniformly from the implants to the bone. Since high interfacial stresses result from the accommodation of a high-modulus prosthesis within the much lower modulus bone, the use of a lower-modulus interpositional material has been introduced as an alternative to PMMA fixation. Thus, polymers such as polysulfone have been used as porous coatings on the implant’s metallic core to permit mechanical interlocking through bone and/or soft tissue ingrowth into the pores. This requires that the polymers have surfaces that resist creep under the stresses found in clinical situations and have high enough yield strength to minimize plastic deformation. As indicated earlier, the mechanical properties of concern in polymer applications are yield stress, creep resistance, and wear rate. These factors are controlled by such polymer parameters as molecular chain structure, molecular weight, and degree of branching. Ceramics and glasses have played an increasingly important role in implants in recent years. This has occurred because of two quite disparate uses. First, there is the use associated with improved properties such as resistance to further oxidation, high stiffness, and low friction and wear in articulating surfaces. This requires the use of full-density, controlled, small, uniform-grain-size (usually less than 5 µm) materials. The small grain size and full density are important since these are the two principal bulk parameters controlling the ceramic’s mechanical properties. Clearly, any void within the ceramic’s body will increase stress, thereby degrading the mechanical properties. Thus, full-density, small-grain-size aluminum oxide, Al2 O3 , has proved quite successful as the material in matched pairs of femoral head and acetabular components in total hip arthoplasty. Other ceramics for this use include TiO2 and Si3 N4 . The second application takes advantage of the osteophilic surface of certain ceramics and glass ceramics. These materials provide an interface of such biological compatibility with bone-forming cells called osteoblasts that these cells lay down bone in direct

MECHANICS OF BIOLOGICS

529

apposition to the material with a direct chemicophysical bond. Special compositions of glass ceramics, termed bioglasses, were originally used for orthopedic implants. The bioglasses have nominal compositions in the ranges 40–50% SiO2 , 5–31% Na2 O, 12–35% CaO, 0–15% P2 O5 , plus in some cases MgO, B2 O3 , or CaF2 . The model proposed for the “chemical bond” formed between the glass and bone is that the former undergoes a controlled surface degradation, producing an SiO2 -rich layer and a Ca, P-rich layer at the interface. This latter layer eventually crystallizes as a mixture of hydroxy-carbonate apatite structurally integrated with collagen, which permits subsequent bonding by newly formed mineralized tissues. There is an entirely different series of ceramic-based compounds that have been shown to be osteophilic. These include hydroxyapatite (HA), which you will recall is the form of the naturally occurring inorganic component of calcified tissues, and calcite, CaCO3 , and its Mg analog, dolomite, among others. The most extensive applications in both orthopedics and dentistry have involved HA. The elastic properties of HA and related compounds such as fluorapatite (FAp), chlorapatite (ClAp), and cobalt apatite (Co3 Ap) are compared with those of bone, dentin, and enamel in Table 5.16. The use of both HA and the glass ceramics as claddings on the metallic stems of prostheses is still another method of providing fixation instead of using PMMA. The fracture toughness properties of these ceramic-based systems are compared to natural bone in Figure 5.132. Composites provide an attractive alternative to the various metal-, polymer- and ceramic-based biomaterials, which all have some mismatch with natural bone properties. A comparison of modulus and fracture toughness values for natural bone provide a basis for the approximate mechanical compatibility required for artificial bone in an exact structural replacement, or to stabilize a bone–implant interface. A precise matching requires a comparison of all the elastic stiffness coefficients (see the generalized Hooke’s Law in Section 5.4.3.1). From Table 5.15 it can be seen that a possible approach to the development of a mechanically compatible artificial bone material

Table 5.16 Selected Mechanical Properties of Various Calcium-Bearing Minerals Compared to Bone and Teeth

Material FAp (mineral) HAp (mineral) HAp (synthetic) ClAp (synthetic Co3 Ap (synthetic, type B) CaF2 Dicalcium phosphate dihydrate Bone (human ferour, powdered) Dentine (bovine, powdered) Enamel (bovine, powdered)

Bulk Modulus (1010 N m−2 )

Shear Modulus (1010 N m−2 )

Elastic Modulus (1010 N m−2 )

Poisson’s Ratio

9.40 8.90 8.80 6.85 8.17 7.74 5.50 2.01 3.22 6.31

4.64 4.45 4.55 3.71 4.26 4.70 2.40 0.800 1.12 2.93

12.0 11.4 11.7 9.43 10.9 11.8 6.33 2.11 3.02 7.69

0.26 0.27 0.28 0.27 0.28 0.26 0.31 0.33 0.35 0.32

Source: L. J. Katz, in Biomaterials Science, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, ed., p. 335. Copyright  1996 by Academic Press.

530

MECHANICS OF MATERIALS

40 Fracture toughness, KIc(MPam1/2)

HA-coated Ti 15

10 Bone 5

HA/bioglass

HA-based ceramic composites

dense HA HA/polymers

0 0

50 100 Elastics modulus (GPa)

150

Figure 5.132 Fracture toughness and modulus of some biomaterials for bone replacement. Reprinted, by permission, from W. Suchanek, and M. Yoshimura, J. Mater. Res., 13(1), 94(1998). Copyright  1998 by Materials Research Society.

is to stiffen one of the polymers already used as an implant with a suitable second phase. The general effect has been well demonstrated for polymers in that a progressive increase in the volume fraction of the second phase will correspondingly increase the elastic modulus but decrease the fracture toughness. Hence, starting with a ductile polymer, an eventual transition into brittle behavior is observed. Because PMMA is brittle, whereas PE is ductile, with an elongation to fracture of about 100%, PE provides the more suitable starting materials and has comparable mechanical behavior to collagen. Various materials could be considered as additions to PE that could stiffen the materials at the expense of fracture toughness, but the selection of HA has proved particularly appropriate. Starting with polyethylene granules of average molecular weight 400,000, along with hydroxyapatite particles of 0.5–20 µm in size, composites of HA-reinforced PE are prepared to various volume fractions by a compounding and molding technique that minimizes polymer degradation. The composites are pore-free, with a homogeneous distribution of HA particles of volume fractions from 0.1 to 0.5. The rule of mixtures applies to the prediction of tensile modulus in these composites, and values increase with increasing HA content up to about 9 GPa at the 50% volume level. The increase in modulus is accompanied by a decrease in the elongation-to-fracture. However, the composite remains unusually ductile and maintains its fracture toughness, up to about 40% volume HA. From 40% to 50% volume fraction, the elongation to fracture decreases rapidly, with brittle fracture observed at Vf = 0.45. At these volume fractions, a fracture toughness of 3–8 MN/m3/2 is observed (cf. Figure 5.132). Hence, in terms of cortical bone replacement, an HA–PE composite with 50% volume HA has a tensile modulus within the lower band of the values for bone and a comparable fracture toughness. Another example of composites in biomedical applications is graphite-fiberreinforced bone cement. Self-curing poly(methyl methacrylate), PMMA, is used extensively as a bone cement in orthopedic surgery for fixation of endoprostheses

MECHANICS OF BIOLOGICS

531

Cooperative Learning Exercise 5.13 A biomedical composite is formed by incorporating 1.05 volume % short carbon fibers into a PMMA matrix. The carbon fibers have the following properties: σf = 2.241 GPa, lc = 2.54 × 10−4 m, l = 1.27 × 10−2 m, Ef = 248.3 GPa, and aspect ratio = 2500. The PMMA has the following properties: σm = 34.48 MPa, Em = 2 GPa. Recall that n = 0.33 for a random fiber composite. Person 1: Use the rule of mixtures to estimate the tensile strength of the composite. Person 2: Use the appropriate equation from Section 5.4.2.2 to estimate the tensile strength of the composite. Compare your answers. Are they the same? Why or why not? How would you go about estimating the modulus for this composite?

Answers: σ1∗ = (0.0105)2241 + (0.9895)34.48 = 57.6 MPa; Eqn. 5.114 for low-volume-fraction fibers: σ1∗ = [2.54 × 10−4 /2(1.27 × 10−2 )](0.0105)(2241) + (0.9895)(34.48) = 34.34 MPa; The corrections for fiber length of Eq. (5.114) introduce a significant and necessary correction. The experimental value is 38.3 MPa. The modulus could be determined from Eq. (5.110), but the product na = 825, which according to Figure 5.95 means β = 1.0, such that a simple rule of mixing can be used for the modulus.

and the repair of bone defects. However, such bone cement is weak in tension and is also significantly weaker than compact bone. The reinforcement of bone cement with graphite fibers has been shown to increase its tensile, compressive, and shear strengths, as well as increase fatigue life and reduce creep. The tensile modulus of fiber-reinforced PMMA also increases with increasing fiber content when the fiber volume percentage 16 000

160 Thornel-300 (1.05 vol% fibers)

Strength (MPa)

120 100

14 000

Thornel-300 (2.08 vol% fibers) 1 wt% aramid fibers 4 wt% aramid fibers

12 000 10 000

human bone 80

8000

60

6000

40

4000

20

2000 Ultimate strength

Elasticity (MPa)

140

Modulus of elasticity

Figure 5.133 Comparison of tensile properties of fiber-reinforced bone cement (PMMA) and human compact bone. Reprinted, by permission, from Concise Encyclopedia of Composite Materials, A. Kelly, ed., p. 270, revised edition. Copyright  1994 by Elsevier Science Publishers, Ltd.

532

MECHANICS OF MATERIALS

is kept small, as illustrated in Figure 5.133, where Thornel is an ultrahigh-strength graphite fiber, and the aramid fiber in use is Kevlar -29. Incorporation of 2% Thornel300 fibers improves the tensile strength of PMMA to 50%, and its modulus to about 40%, of that of compact human bone. As described in Section 5.4.2.7, the energy absorption capacity of the composite is increased due to fiber pull-out. REFERENCES Cited References 1. Li, X., and B. Bhushan, A review of nanoindentation continuous stiffness measurement technique and its applications, Mater. Char., 48, 11–36 (2002). 2. Jenkins, C. H. M., The strength of Cd–Zn and Sn–Pb alloy solder, J. Inst. Metals, 40, 21–39 (1928). 3. Pearson, C. E., The viscous properties of extruded eutectic alloys of lead-tin and bismuth-tin, J. Ins. Metals, 54, 111–124 (1934). 4. Bochavar, A. A., and Z. A. Sviderskaya, Superplasticity in Zn–Al alloys, Bull. Acad. Sci. (USSR) 9, 821–824 (1945). 5. Underwood, E. E., A review of superplasticity, JOM, 914–919 (1962). 6. Holt, D. L., and W. A. Backofen, Superplasticity in the Al–Cu eutectic alloy, Trans. ASM, 59, 755–767 (1966). 7. Ashby, M. F., Materials Selection in Mechanical Design, 2nd ed., Butterworth Heinemann, Oxford, 1992. 8. Concise Encyclopedia of Composite Materials, A. Kelly, ed., Pergamon, Elmsford, NY, 1994, p. 100. 9. Concise Encyclopedia of Composite Materials, A. Kelly, ed., Pergamon, Elmsfords NY, 1994, p. 161. 10. Powell, P. C., and A. J. Ingen Housz, Engineering with Polymers, Stanley Thornes, Cheltenham, UK, 1998. 11. Suchanek, W., and M. Yoshimura, Processing and properties of hydroxyapatite-based biomaterials for use as hard tissue replacement implants, J. Mater. Res., 13(1), 94 (1998).

Mechanics of Materials Knott, J., and P. Withey, Fracture Mechanics: Worked Examples, Institute of Materials, London, 1993.

Mechanical Properties of Metals Metals Handbook, Desk Edition, American Society for Metals, 1985.

Mechanical Properties of Ceramics and Glasses Wachtman, J. B., Mechanical Properties of Ceramics, John & Sons, Wiley, New York, 1996.

Mechanical Properties of Polymers McCrum, N. G., C. P. Buckley, and C. B. Bucknall, Principles of Polymer Engineering, Oxford Science, Oxford, 1997.

PROBLEMS

533

Mechanical Properties of Composites Kelly, A., Composite Materials, American Elsevier, New York, 1966. Composite Materials Technology, P. K. Mallick and S. Newman, eds., Hanser Publishers, Munich, 1990. Mechanical Testing of Engineering Ceramics at High Temperatures, B. F. Dyson, R. D. Lohr, and R. Morrell, ed., Elsevier Applied Science, London, 1989. Lee, W. E., and W. M. Rainforth, Ceramic Microstructures: Property Control by Processing, Chapman & Hall, London, 1994. Composite Applications: The Role of Matrix, Fiber and Interface, T. L. Vigo and B. J. Kinzig, eds., VCH Publishing, New York, 1992.

Mechanical Properties of Biologics The mechanical properties of biological materials, Symposium, Society of Experimental Biologists, Cambridge University Press, New York, 1980. Biomaterials Properties Database, University of Michigan, Dr. William J. O’Brien, Director http://www.lib.umich.edu/dentlib/Dental tables/toc.html Evans, F. E., Mechanical Properties of Bone, C. Charles, Springfield, IL, 1973.

PROBLEMS Level I

5.I.1 The following three stress–strain data points are provided for a titanium alloy: ε = 0.002778 at σ = 300 MPa; ε = 0.005556 at σ = 600 MPa; ε = 0.009897 at σ = 900 MPa. Calculate the modulus of elasticity for this alloy. 5.I.2 Experimentally, it has been observed for single crystals of a number of metals that the critical resolved shear stress is a function of the dislocation density, ρD : τcr = τ0 + A(ρD )0.5 where τ0 and A are constants. For copper, the critical resolved shear stress is 2.10 MPa at a dislocation density of 105 mm−2 . If it is known that the value of A for copper is 6.35 × 10−3 MPa·mm, compute the critical resolved shear stress at a dislocation density of 107 mm−2 . 5.I.3 Calculate the composite modulus for epoxy reinforced with 70 vol% boron filaments under isotropic strain conditions. 5.I.4 A single crystal of aluminum is oriented for a tensile test such that its slip plane normal makes an angle of 28.1◦ with the tensile axis. Three possible slip directions make angles of 51.2◦ , 36.0◦ , and 40.6◦ with the same tensile axis. Which of these three slip directions is most favored? 5.I.5 According to the Kelvin (Voigt) model of viscoelasticity, what is the viscosity (in Pa·s) of a material that exhibits a shear stress of 9.32 × 109 Pa at a shear strain of 0.5 cm/cm over a duration of 100 seconds? The shear modulus of this material is 5 × 109 Pa.

534

MECHANICS OF MATERIALS

5.I.6 A high-strength steel has a yield strength of 1460 MPa and a K1c of 1

98 MPa·m 2 . Calculate the size of a surface crack that will lead to catastrophic failure at an applied stress of 21 σy . 5.I.7 An alloy is evaluated for potential creep deformation in a short-term laboratory experiment. The creep rate is found to be 1% per hour at 880◦ C and 5.5 × 10−2 % per hour at 700◦ C. (a) Calculate the activation energy for creep in this temperature range. (b) Estimate the creep rate to be expected at a service temperature of 500◦ C. (c) What important assumption underlies the validity of your answer to part b? 5.I.8 A glass plate contains an atomic-scale surface crack. Assume that the crack tip radius is approximately equal to the diameter of an O2− ion. Given that the crack is 1 µm long and the theoretical strength of the defect-free glass is 7.0 GPa, calculate the breaking strength of the glass plate. Level II

5.II.1 In normal motion, the load exerted on the hip joint is 2.5 times body weight. (a) Estimate the corresponding stress (in MPa) on an artificial hip implant with a cross-sectional area of 5.64 cm2 in an average-sized patient, and (b) calculate the corresponding strain if the implant is made of Ti–6Al–4V alloy. Clearly state your assumptions and sources of data. 5.II.2 In the absence of stress, the center-to-center atomic separation distance of two Fe atoms is 0.2490 nm along the [111] direction. Under a tensile stress of 1000 MPa along this direction, the atomic separation distance increases to 0.1489 nm. (a) Calculate the modulus of elasticity along the [111] direction. (b) How does this compare with the average modulus for polycrystalline iron with random grain orientation? (c) Calculate the center-to-center separation distance of two Fe atoms along the [100] direction in unstressed α-iron. (d) Calculate the separation distance along the [100] direction under a tensile stress of 1000 MPa. 5.II.3 It is sometimes convenient to express the degree of plastic deformation in metals as the percent cold work, CW%, rather than as strain: %CW = [(A0 − Af )/A0 ] × 100 where A0 and Af are the initial and final cross-sectional areas, respectively. Show that for a tensile test %CW = [ε/(ε + 1)] × 100 Look up the stress–strain diagram for naval brass and determine the CW% where a stress of 400 MPa is applied. Cite the source of your information. Level III

5.III.1

In a recent article [Shinzato, S., et al., A new bioactive bone cement: Effect of glass bead filler content on mechanical and biological properties, J. Biomed.

535

PROBLEMS

Mater. Res., 54(4), 491 (2000)] the mechanical properties of glass-bead-filled poly(methylmethacrylate) (PMMA) bone cement are described. Four plots from this article are reproduced below. “GBC” refers to the glass bead composite, and the number refers to the weight percent glass beads; for example, GBC50 has 50 wt% glass beads in the PMMA matrix. The data from all tests were obtained from rectangular samples, 20 × 4 × 3 mm (length × width × thickness). (a) Calculate the radius (in mm) of the cylindrical samples used for the compressive tests. Note that the compressive strengths of samples GBC30, GBC40, and GBC50 shown in Figure (a) were taken as the inflection point in the stress–strain diagrams shown in Figure (d), rather than the break point, as in samples GBC60 and GBC70. (b) Estimate Poisson’s ratio for this material. (c) What is the force (in terms of weight, kg) required to break the GBC samples in a three-point bending test? (d) Sketch the stress–strain diagrams that result from the three-point bend test for the following three materials: glass beads (assume they are a generic silicate glass), PMMA, and GBC50. You should have three separate curves, each representing the relative behaviors of the three materials.

Compressive Strength

Bending Strength 160

160

140

140 119

100 80

88

89

99

60

74

136

135

133

134

100 92

80 60

40

40

20

20

0

137

120

131 (MPa)

(MPa)

120

0

GBC30 GBC40 GBC50 GBC60 GBC70 PMMA

GBC30 GBC40 GBC50 GBC60 GBC70 PMMA (b)

(a)

Stress-strain Curves for Compressive Tests (Kgf) 500 GBC70 400

GBC60

GBC50 GBC40 GBC30

Stress

300 Young's Modulus 9 8

8.4

7

200 : Break point

(GPa)

6 5.8

5 4 3

3.6

4.1

: Point of inflexion

4.6

2

100 2.7

1 0

GBC30 GBC40 GBC50 GBC60 GBC70 PMMA (c)

0

1

2 Strain (d)

3

4 (mm)

5.III.2

MECHANICS OF MATERIALS

In another article on bone cements [Vallo, C, Influence of filler content on static properties of glass-reinforced bone cement, J. Biomed. Mater. Res., 53(6), 717 (2000)], the following data for the critical intensity factor in these same glass-bead-filled PMMA composites as was used in Problem 5.III.1 are given: 3.0

2.5

KIC (MPa m1/2)

536

2.0

1.5

Cured at room temperature 1.0 3.0 0

10

20 30 Glass wt%

40

50

Using the values of the compressive strength from Problem 5.III.1 as representative of the stress at the fracture instability (critical stress for crack propagation), estimate the crack length at the fracture instability, a (in mm), for the GBC material. Clearly indicate which composition you are using in your calculation and any assumptions you must make.

CHAPTER 6

Electrical, Magnetic, and Optical Properties of Materials

This chapter is organized much like Chapter 4. We will describe three related properties of materials, and we will subdivide each section into the five materials classes. Just as in Chapter 4, there are some fundamental equations that relate important fluxes to driving forces and, in doing so, create proportionality constants that are material properties. And, just as in Chapter 4, we will be more concerned with the constants—material properties—than with the driving forces. In fact, we will be even less concerned with the driving forces and fluxes of electricity and magnetism in this chapter than we were with momentum, heat, and mass transport topics, since the latter have more to do with materials processing, which we will touch upon in Chapter 7, whereas the former are more the realm of solid-state physics. This is not to say that neither the origins of the driving forces nor their resultant fluxes are unimportant to the engineer. On the contrary, these are the design parameters that must often be met. The point is that the derivations of many of these fundamental equations will be dispensed with in favor of concentration upon their application. The third topic of this chapter is related to the first two, although no fundamental equation of the driving-force form exists. It has to do with how electromagnetic radiation interacts with materials, and it is a logical extension of the description of electrical and magnetic phenomena. The first topic will deal with electrical conductivity. The driving force for the flow of electrons that leads to electrical conduction is the electric field, E, and the resulting flow of electrons is characterized by the current density, J. The flow and driving force are related through Ohm’s Law: J = σE (6.1) where J is the current density in units of amps/m2 , E is the electric field strength in units of volts/m, and σ is the electrical conductivity in units of ohms−1 /m or ( · m)−1 . It is indeed unfortunate that the symbol for conductivity is also the symbol we utilized for the strength of a material, but it is rare that the two are used in the same context, so that confusion should be easily avoided. The second section will address the topic of magnetism. The magnetic driving force is the magnetic field, H, and is related to a quantity known as the magnetic flux density, or magnetic induction, B, through the magnetic analogue to Ohm’s Law: B = µH

(6.2)

An Introduction to Materials Engineering and Science: For Chemical and Materials Engineers, by Brian S. Mitchell ISBN 0-471-43623-2 Copyright  2004 John Wiley & Sons, Inc.

537

538

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

where B has units of teslas (webers/m2 ), and H has units of amps/m. The proportionality constant, µ, is called the magnetic permeability. We will have much more to say about the units of permeability in Section 6.2, since there is some variability in its definition. It is once again unfortunate that the symbol for magnetic permeability is identical to that of another important engineering parameter, Newtonian viscosity, but there is little opportunity for confusion in real applications. The third topic has to do with how materials interact with radiation that has both an electrical and a magnetic component—that is, electromagnetic radiation. This topic is often generically referred to as “optical properties” of materials, even though the full range of electromagnetic radiation, of which the optical portion is only a small segment, can be considered. There is no unifying, driving-force-containing equation to describe the optical properties of materials. However, there is one useful relationship that we will use to build most of the descriptions of Section 6.3 upon which relates the intensity of incident electromagnetic radiation, I0 , to the transmitted, absorbed, and reflected intensities, IT , IA , and IR , respectively: I0 = IT + IA + IR

(6.3)

By the end of this chapter, you should be able to: ž ž ž ž ž ž ž ž ž ž ž ž ž ž ž ž

6.1

Relate conductivity, resistivity, resistance, and conductance. Calculate conductivity from charge carrier concentration, charge, and mobility. Differentiate between a conductor, insulator, semiconductor, and superconductor. Differentiate between an intrinsic and an extrinsic semiconductor. Differentiate between a p- and n-type semiconductor. Identify different types of electrical polarizability, and determine if they are relevant to a given chemical structure. Describe how a p –n semiconductor junction works. Differentiate between diamagnetism, paramagnetism, ferromagnetism, ferrimagnetism, and antiferromagnetism. Describe the temperature dependence of magnetic susceptibility. Define the Curie temperature for a substance. Define the Ne´el temperature for a substance. Determine coercivity and remnant induction from a hysteresis loop. Relate transmissivity, absorptivity, and reflectivity. Define refractive index. Calculate reflectivity from refractive index differences at an interface. Describe the interactions of light with materials that result in color. ELECTRICAL PROPERTIES OF MATERIALS

Without going into a complete derivation, let us examine Eq. (6.1) more closely to see where the driving force, electrical current, and electrical conductivity come from. The electric field, E, is a vector quantity, as indicated by the boldfaced type. For the time-being, we will limit the description to an electric field in one direction only and

ELECTRICAL PROPERTIES OF MATERIALS

539

Variable resistor

Ammeter

I

Battery

V

Cross-sectional area, A

Specimen

Voltmeter

Figure 6.1 Schematic diagram of conductivity measurement in a material. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 607. Copyright  2000 by John Wiley & Sons, Inc.

will consider only the magnitude of the electric field, E . An electric field is generated by applying a voltage, V , across a distance, l, as illustrated in Figure 6.1. The electric field strength is then given by E = V/l (6.4) When a conductive material is placed within the electric field, current begins to flow, as characterized by the current density, J, of Eq. (6.1). The current density is also a vector quantity, but since our field is in one dimension only, current will similarly flow only in one direction, so that we will use only the scalar quantity from here on, J . The current density is simply the current, I, per unit area in the specimen, A: I A

J =

(6.5)

where I is in units of ampere (amps, A, for short) and A is in units of m2 . The current, in turn, is defined as the rate of charge passage per unit time: I =

dq dt

(6.6)

where q is the charge in units of coulombs, so that an ampere is equal to a coulomb/s. Substitution of Eqs. (6.4) and (6.5) into Eq. (6.1) and isolation of V gives V =I



l σA



(6.7)

We define the quantity in parentheses as the resistance, R, to obtain the more familiar form of Ohm’s Law: V =I ×R (6.8)

540

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Cooperative Learning Exercise 6.1

Consider a wire, 3 mm in diameter and 2 m long. Use the data in Appendix 8 to answer the following questions. Person 1: What is the resistance in the wire if it is made of copper? Person 2: What is the resistance in the wire if it is made of stainless steel? Exchange the results of your calculations, check them, and use them to make the following calculations, assuming that a potential difference of 0.05 V is placed across the ends of the wire. Person 1: If the wire is made of stainless steel, calculate the current flow, current density, and magnitude of the electric field. Person 2: If the wire is made of copper, calculate the current flow, current density, and magnitude of the electric field. Answers: RCu = 4.84 × 10−3 ; RSS = 0.202 .ISS = 0.25 A, JSS = 35,000 A/m2 , ESS = 0.025 V/m; ICu = 10.3 A; JCu = 1.5 × 106 A/m2 , ECu = 0.025 V/m.

The relationship between conductivity and resistance is an important one, and it can be further generalized through the introduction of two additional quantities called the conductance, G, and the resistivity, ρ. The resistivity is simply the resistance per unit length times the cross-sectional area, ρ = RA/ l, and as such is a geometryindependent material property, as is the conductivity, σ . Similarly, the conductance is the conductivity per unit area times length, G = σ l/A. It is important to remember that the geometry of the sample must be known to calculate its resistance and/or conductance, whereas the resistivity and conductivity are inherent material properties that can vary with temperature and composition, but are generally independent of sample geometry. Moreover, it is useful to keep in mind that conductivity and resistivity are inverses of one another, and if we know one, we can find the other. Conductivity requires a charge carrier. There are two types of charge carriers we will consider: electrons and ions. The structural descriptions of Chapter 1 will be helpful in determining the primary type of charge carrier within a material, if any. In subsequent sections, we explore the molecular origins of each type of conductivity, investigate the important parameters that cause conductivity to vary in materials, and describe additional electrical conduction phenomena that have revolutionized our lives. 6.1.1

Electrical Properties of Metals and Alloys

Electrical conduction in metals and alloys occurs by the motion of electrons. It can be shown that the conductivity is proportional to the number of electrons per unit volume, ne , the charge per electron, qe , and the electron mobility, µe : σ = ne qe µe

(6.9)

where the units of ne are m−3 , qe is the magnitude of charge on an electron (1.602 × 10−19 coulombs), and µe is in m2 /V · s. Any of these three quantities can, and do, affect the overall conductivity of a material, leading to an astounding 27 orders of magnitude in the variation of electrical conductivity values. This immense variation

ELECTRICAL PROPERTIES OF MATERIALS

541

in a physical property causes us to further classify materials according to their ability to conduct electricity. Those materials for which the value of σ is low, ranging from 10−20 to 10−10 ( · m)−1 , are termed insulators, while those with intermediate values on the order of 10−6 to 104 ( · m)−1 are called semiconductors. Materials with high electrical conductivities above 107 ( · m)−1 , such as most metals and alloys, are called conductors. Notice that the range of conductivities for these classes are still very large and that there are significant gaps between them for which absolute classification is difficult. In fact, a given material can cross over from one classification to another, depending on such factors as temperature. As a result, it is important to understand the origins of electrical conduction in these materials. 6.1.1.1 The Molecular Origins of Electrical Conduction. Recall from Section 1.0.4.3 that the bonds in metals tend to form bands of electrons. The electrons involved in bonding are found in the valence band, inner electrons not involved in bonding are found in the core band, and the remaining unfilled orbitals of the outer bands form the conduction band. It is the conduction band that gives metals and alloys the ability to freely conduct electrons. As shown in Figure 6.2, the “distance” between the conduction and valence bands in a solid, called the band gap, Eg , can be used to characterize the electrical behavior of a material. Insulators have large band gaps of several electron volts (eV), whereas conductors have small band gaps, or even overlapping valence and conduction bands, as illustrated in Figure 6.2b. Metallic solids are good electrical conductors because their valence band is not completely filled or because it may overlap with a second empty valence band, thereby causing the total energy band to be only partially filled. According to the band theory, the motion of electrons under an external electric field is possible only for electrons in partially filled energy bands. The probability function,

Conducting band Conducting band

Energy levels

Energy gap

Eg Valence band

Valence band

Core band (a)

Core band (b)

Figure 6.2 Schematic illustration of band gaps in (a) an insulator and (b) conductor. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

542

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

f (E), that describes the probability that an energy level E is occupied by an electron is given by Fermi as 1 (6.10) f (E) = exp[(E − Ef )/kB T ] + 1 where Ef is called the Fermi energy, kB is Boltzmann’s constant, and T is the absolute temperature. The Fermi energy in a metal represents the kinetic energy of the electrons having the maximum energy level that some electrons can attain. At absolute zero, the Fermi–Dirac distribution law requires that the Fermi function, f (E), must have a value of 1 or 0 (see Figure 6.3a). Thus, at T = 0 K we have f (E) = 1 and E < Ef . At temperatures greater than 0 K, f (E) changes from 1 to 0 over an energy range of about one unit of kB T . At E = Ef we obtain f (E) = 21 (see Figure 6.3b). If E − Ef >> kB T , the exponential term in the denominator of Eq. (6.10) becomes very large relative to unity, and f (E) becomes 

−(E − Ef ) f (E) = exp kB T



(6.11)

1

1

f (E )

f (E )

T = 300 K

T=0K

0.5

T = 1000 K

0

0

Ef

Energy

E0 – 0.1 eV Ef E0 + 0.1 eV Energy

(a)

(b)

N (E )

T=0 Unfilled levels T≠0

Filled levels

E

Ef

E max

(c)

Figure 6.3 The Fermi distribution function (a) at absolute zero and (b) at a finite temperature. (c) The population density of electrons in a metal as a function of energy. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

543

For electron energies, E, below the Fermi level, Ef − E >> kB T , and Eq. (6.11) can be approximated as   E − Ef (6.12) f (E)≈1 − exp kB T indicating that the probability of occupation of the energy band is nearly unity. The state densities are not uniform across the energy band, and their population density, N (E), is the greatest in the center of the band. The number of electrons in the band, ne , can then be evaluated and used in Eq. (6.9) by integrating the product of the density of state N (E) and the probability of their occupation, f (E), over the band energy range (see Figure 6.3c). Thus, for metals, ne =



Ef 0

N (E) · f (E) dE

(6.13)

The energy distribution in the conduction zone extends over several electron volts and is about 5 eV for some metals at absolute zero. Thus, only a small fraction of the electrons in the energy band can be excited above the Fermi level, and only those within an energy range of the order kB T can be excited thermally. In addition to the number of electrons, the other factor in Eq. (6.9) that affects conductivity is the electron mobility, µe . The mobility is the average charge carrier velocity, or drift velocity, v, divided by the electric field strength, E : µe = v/E

(6.14)

In a typical metal, µe = 5 × 10−3 m2 /V · s, which gives a drift velocity of 5 × 10−3 m/s for an electric field of 1 V/m. The average electron velocity, in turn, is related to an important structural parameter, the mean free path, l, which we utilized in Section 4.2.1.4 to describe thermal conductivity: l = τv

(6.15)

where τ is a quantity known as the relaxation time and has units of sec−1 . The relaxation time is closely related to the mean time of flight between electron collisions and also to the mean free path of the conduction electrons, as indicated above. As the mobilities are likely to depend on temperature only as a simple power law over an appropriate region, the temperature dependence on conductivity will be dominated by the exponential dependence of the carrier concentration. We will have more to say about carrier mobility in the section on semiconductors. In summary, metals are good electrical conductors because thermal energy is sufficient to promote electrons above the Fermi level to otherwise unoccupied energy levels. At these levels (E > Ef ), the accessibility of unoccupied levels in adjacent atoms yields high mobility of conduction electrons known as free electrons through the solid. 6.1.1.2 Resistivity in Metals and Alloys. Resistivity is the reciprocal of conductivity. We will find it useful to momentarily concentrate on resistivity instead of conductivity to explain some important electrical phenomena. If the metal crystal lattice were perfect and there were no lattice vibrations, the electrons would pass through

544

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

the lattice unscattered, encountering no resistance. But, as we saw in the description of thermal conductivity (cf. Section 4.2.2.1), lattice vibrations and phonon scattering play a role in disrupting the mean free path of electrons. Working backwards through the previous equations, then, shows that a reduction in mean free path, l [Eq. (6.15)], due to phonon scattering and lattice vibrations decreases the electron mobility [Eq. (6.14)], which in turn can reduced electrical conductivity [Eq. (6.9)] or increase resistivity. This effect leads to the temperature-dependent portion of resistivity, ρt . Since the thermal lattice vibrations increase directly with increasing temperature, the electrical resistivity of metals will increase with temperature in a linear manner. At high temperatures approaching the melting point, the resistivity usually rises somewhat more rapidly than linearly with temperature. Only at very low temperatures can lattice defects and impurity atoms affect conductivity and resistivity, rather than thermal lattice vibrations. This type of resistivity is known as the residual resistivity, ρr , and is relatively temperatureindependent. The two types of resistivity are additive, such that the total resistivity is given by ρ = ρt + ρr (6.16) This relationship is shown in Figure 6.4 for a series of copper alloys. The variation of resistivity with temperature is approximately linear over a wide temperature range, and near absolute zero the alloys have a relatively temperature-independent residual resistivity. Notice that as the level of nickel impurities increases, so do both the residual

6

5 Cu – 3.32% Ni

r (10–8 Ω • m)

4 Cu – 2.16% Ni 3

Cu – 1.12% Ni

2

1

0

Cu

0

100

200

300

T (K )

Figure 6.4 Variation of resistivity with temperature for several copper alloys. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

545

and thermal vibration resistivities. This is directly related to a reduction in the electron mean free path. Oftentimes the variation of resistivity with temperature in the linear region is approximated with an empirical expression of the form ρt = ρ0 + aT

(6.17)

where ρ0 and a are experimentally determined constants for a particular metal. Do not confuse the empirical parameter ρ0 with the residual resistivity, ρr - the former is an equation-fitting parameter only for the linear region of the resistivity-temperature plot, whereas the latter is a fundamental property of a material. The resistivities of some metals and alloys are presented in Appendix 8. The variation of resistivity with composition is also expressed in an empirical fashion. For two-phase alloys consisting of phases α and β, the rule of mixtures can be used to approximate the alloy resistivity from the individual metal resistivities: ρ = Vα ρα + (1 − Vα )ρβ

(6.18)

where Vα is the volume fraction of phase α. Other factors such as cold deformation and processing conditions can affect resistivity. The cold deformation effect is not as pronounced as the addition of impurity or alloying elements, but the effect of processing on resistivity can be large, as illustrated in Figure 6.5 for a Cu3 Au compound. The effect on resistivity is again related to structure. The rapidly quenched compound maintains its disordered structure, which

20

r (10–8 Ω • m)

15

d Quenche ered) rd o is (d u Cu 3A

10

d coole Slowly ordered) u( Cu 3A

5

0

0

100

Order – disorder transformation temperature

200

300

400

500

T (°C)

Figure 6.5 Temperature variation of the resistivity for Cu3 Au under two different processing conditions. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

546

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

causes an increase in the resistivity relative to the slowly cooled, ordered structure. In the ordered structure, gold atoms occupy the corner positions and copper atoms occupy the face-centered positions of the FCC unit cell, which results in less electron scattering than the disordered structure in which copper and gold atoms occupy FCC atomic sites at random within each unit cell. Only at the order–disorder transition temperature do the two resistivity curves become coincident, corresponding to a transformation from the ordered to the disordered FCC structure as a result of increased thermal motion. For metals in general, any mechanical or chemical action that alters the crystalline perfection will raise the residual resistivity and, therefore, the total resistivity, according to Eq. (6.16). Thus, vacancies in metals, in contrast to those in ionic solids, increase the resistivity. The reason for this lies in the inherent differences between conduction mechanisms in these two classes of materials. The differences between ionic and electronic conduction will be elaborated upon in Section 6.1.2. 6.1.1.3 Superconduction. You may have noticed in Figure 6.3 that the resistivity of pure copper approaches zero at absolute zero temperature; that is, the residual resistivity is zero. An expanded scale shows that this is not really the case. Figure 6.6 shows that the residual resistivity in pure copper is about 10−10  · m. This is the “normal” behavior of many metals. Some metals, however, such as Pb, lose all resistivity abruptly and completely at some low temperature. This phenomenon is called superconductivity, and the materials that exhibit it are called superconductors. The temperature at which the resistivity vanishes is called the critical transition temperature, Tc . For example, lead has a Tc = 7.19 K, and it has a resistivity that is less than 10−25  · m at 4.2 K. Many elemental metals (see Figure 6.7), solid–solution alloys, and intermetallic compounds exhibit superconductivity. Often, such materials have relatively high electrical resistivities above Tc , compared to other nonsuperconducting metals such as Ag and Cu. Observed Tc values range from a few millidegrees Kelvin to over 21 K for intermetallic compounds such as Nb3 Ge. In 1957 the research team of Bardeen, Cooper, and Schrieffer produced a theory, now known as the BCS theory, that managed to explain all the major properties of

HISTORICAL HIGHLIGHT

The discovery of superconductivity was not very dramatic. When Dutch physicist Heike Kamerlingh Onnes succeeded in liquefying helium in 1908 he looked around for something worth measuring at that temperature range. His choice fell upon the resistivity of metals. He tried platinum first and found that its resistivity continued to decline at lower temperatures, tending to some small but finite value as the temperature approached absolute zero. He could have tried a large number of other metals with similar prosaic results. But he was in luck. His second metal, mercury, showed quite unorthodox behavior, and in 1911 he showed that its resistivity suddenly

decreased to such a small value that he was unable to measure it—and no one has succeeded in measuring it ever since. In the subsequent years, many more superconductors were found at these very low temperatures. By the 1960s, certain alloys of niobium were made that became superconductors at 10–23 K. It was generally believed on theoretical grounds that there would be no superconductors above 30 K. See Section 6.1.2.4 to see if this belief is still held today. Source: Adapted from Solymar and Walsh [1] and Sheahan [2].

ELECTRICAL PROPERTIES OF MATERIALS

547

10–7

r (Ω • m)

10–8

Pb (a superconductor)

10–9

10–10 Cu (a normal metal)

10–11 0 10

102

Tc 101 T (K)

Figure 6.6 Variation of resistance at low temperatures for a nonsuperconductor and a superconductor. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

VIIA

IA IIIA

IIA 4 Be IIIB

IVB VB VIB VIIB 22 Ti 40 Zr

23 V 41 Nb 73 Ta

42 Mo 74 W

43 Tc 75 Re

VIII

44 Ru 76 Os

77 Ir

IB

13 IIB Al 30 31 Zn Ga 48 49 Cd In 80 Hg

IVA VA VIA VIIA

50 Sn 82 Pb

51 Sb

57 La 90 Th

91 Pa

Figure 6.7 The superconducting elements. Reprinted, by permission, from I. Amato, Stuff, p. 68. Copyright  1997 by Ivan Amato.

548

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

superconductivity. The essence of the theory is that superconductivity is caused by electron–lattice interaction and that the superconducting electrons consist of paired ordinary electrons called a Cooper pair. We give here a qualitative description of how Cooper pairs form and how they lead to superconductivity. Figure 6.8a shows the energy–momentum curve of an ordinary conductor with seven electrons in their discrete energy levels. In the absence of an electric field, the current from electrons moving to the right is exactly balanced by the electrons moving to the left, and the net current is zero. When an electric field is applied, all the electrons acquire some extra momentum, and this is equivalent to shifting the whole distribution in the direction of the electric field, as shown in Figure 6.8b. When the electric field is once again removed, the faster electrons will be scattered into lower energy states due to collisions with the vibrating lattice or impurity atoms, and the original distribution in Figure 6.8a is reestablished. For the model of Figure 6.8b, it means that the electron is scattered from energy level a to energy level b in order to reestablish the original distribution. In the case of a superconductor, it becomes energetically more favorable for the electrons to pair up. Those of opposite momenta pair up to form a new particle called a superconducting electron or Cooper pair. This link between two electrons is shown in Figure 6.9a by an imaginary spring. The velocity of the entire center of mass is zero. According to the de Broglie relationship (λ = h/p, where λ is the wavelength, h is Planck’s constant, and p is the momentum), this means that the wavelength associated with the new particle, λ, is infinitely long. This is valid for all the Cooper pairs. So, there is a large number of identical particles all with infinite wavelength—that is, a quantum phenomenon on a macroscopic scale. An applied electric field will displace all the particles again as shown in Figure 6.9b, but this time, when the electric field disappears, there is no return to the previous state. Scattering from energy level a to energy level b is no longer possible because then the electrons at both b and c would become pairless, which is energetically unfavorable. So, the asymmetrical distribution remains and there will be more electrons going to the right than to the left, and the current persists, in theory, forever. As the temperature rises, thermal energy eventually overcomes the pairing interaction, and the superconductivity effect vanishes.

Energy

Energy

a

0 (a)

Momentum

b

0 (b)

Momentum

Figure 6.8 One dimensional schematic representation of the energy momentum curve for seven electrons in a conductor with (a) no applied electric field and (b) an applied electric field. Reprinted, by permission, from L. Solymar, and D. Walsh, Lectures on the Electrical Properties of Materials, 5th ed., p. 427. Copyright  1993 by Oxford University Press.

549

Energy

ELECTRICAL PROPERTIES OF MATERIALS

Momentum

Energy

0 (a)

a

b c

0 (b)

Momentum

Figure 6.9 The energy–momentum curve for seven electrons in a superconductor with (a) no applied electric field and (b) an applied electric field. Reprinted, by permission, from L. Solymar and D. Walsh, Lectures on the Electrical Properties of Materials, 5th ed., p. 428. Copyright  1993 by Oxford University Press.

One technologically significant application for superconductivity is in the use of electromagnets. It would be beneficial to have high magnetic fields without any power dissipation. Early experimenters, however, found that above certain magnetic fields, the superconductor became normal, regardless of the temperature. Thus, in order to have zero resistance, a superconductor must be held not only below the critical transition temperature, Tc , but also below a critical magnetic field, Hc . The interrelation between Hc and Tc is well described by the formula 

Hc = H0 1 −



T Tc

2 

(6.19)

where H0 is the critical magnetic field at absolute zero. This relationship is graphically illustrated in Figure 6.10, where the material is a “normal” conductor above the curve and a superconductor below the curve. The values of H0 and Tc for a number of elemental superconductors are given in Table 6.1. We will have more to say about superconductors in Section 6.1.2.4.

550

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

H H0 Normal state

Superconducting state

T

Tc

0

Figure 6.10 Temperature dependence of critical magnetic field for a superconductor. Reprinted, by permission, from L. Solymar and D. Walsh, Lectures on the Electrical Properties of Materials, 5th ed., p. 429. Copyright  1993 by Oxford University Press.

Table 6.1 Critical Temperature and Critical Magnetic Field of a Number of Superconducting Elements Element Al Ga Hg α Hg β In Nb

Tc (K)

H0 × 10−4 (A/m)

1.19 1.09 4.15 3.95 3.41 9.46

0.8 0.4 3.3 2.7 2.3 15.6

Element Pb Sn Ta Th V Zn

Tc (K)

H0 × 10−4 (A/m)

7.18 3.72 4.48 1.37 5.30 0.92

6.5 2.5 6.7 1.3 10.5 0.4

Source: L. Solymar, and D. Walsh, Lectures on the Electrical Properties of Materials, 5th ed. Copyright  1993 by Oxford University Press.

6.1.1.4 Intrinsic Semiconduction. We return now to our description of band gaps from Section 6.1.1.1 in order to elaborate upon those materials with conductivities in the range 10−6 to 104 ( · m)−1 . At first, it may not seem important to study materials that neither readily conduct electricity nor protect against it, as do insulators. However, the development of semiconducting materials has, and continues to, revolutionize our lives. Recall from Figure 6.2 that the gap between the valence and conduction bands called the band gap, Eg , can be used to classify materials as conductors, insulators, or semiconductors; also recall that for semiconductors the value of Eg is typically on the order of 1–2 eV. The magnitude of the band gap is characteristic of the lattice alone and varies widely for different crystals. In semiconductors, the valence and conduction bands do not overlap as in metals, but there are enough electrons in the valence band that can be “promoted” to the conduction band at a certain temperature to allow for limited electrical conduction. For example, in silicon, the energies of the valence electrons that bind the crystal together lie in the valence band. All four

ELECTRICAL PROPERTIES OF MATERIALS

T=0K

551

T = 300 K Conduction band

Empty

Ec

Energy

Eg Ef

Valence band

+

+

+

+

Ev

Full

Figure 6.11 Energy bands of an intrinsic semiconductor: Ef is the Fermi energy level; Ec is the lower edge of the conduction band; Ev is the upper edge of the valence band; and Eg is the band gap. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

valence electrons for each atom are tied in place, forming covalent bonds. There are no electrons in the conduction band at absolute zero, and the material is effectively an insulator (see Figure 6.11). At an elevated temperature, enough electrons have been thermally promoted to the conduction band to permit limited conduction. Promotion of electrons leaves behind positively charged “holes” in the valence band which maintain charge neutrality. The term hole denotes a mobile vacancy in the electronic valence structure of a semiconductor that is produced by removing one electron from a valence band. Holes, in almost all respects, can be regarded as moving positive charges through the crystal with a charge magnitude and mass the same as that of an electron. Conduction that arises from thermally or optically excited electrons is called intrinsic semiconduction. The conduction of intrinsic semiconductors usually takes place at elevated temperatures, since sufficient thermal agitation is necessary to transfer a reasonable number of electrons from the valence band to the conduction band. The elements that are capable of intrinsic semiconduction are relatively limited and are shown in Figure 6.12. The most important of these are silicon and germanium. In intrinsic semiconductors, the number of holes equals the number of mobile electrons. The resulting electrical conductivity is the sum of the conductivities of the valence band and conduction band charge carriers, which are holes and electrons, respectively. In this case, the conductivity can be expressed by modifying Eq. (6.9) to account for both charge carriers: σ = ne qe µe + nh qh µh

(6.20)

where nh , qh , and µh are now the number, charge, and mobility of the holes, respectively, and contribute to the overall conductivity in addition to the corresponding

552

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

VIIA

IA IIA

IIIB IVB VB VIB VIIB —

VIII



IB

IIB 30 Zn 48 Cd 80 Hg

IIIA IVA VA VIA VIIA 8 O 13 14 15 16 Al Si P S 31 32 33 34 Ga Ge As Se 49 50 51 52 In Sn Sb Te

Figure 6.12 The intrinsic semiconducting elements. Reprinted, by permission, from I. Amato, Stuff, p. 68. Copyright  1997 by Ivan Amato.

quantities for the electrons, indicated by the subscript e. The charge of the holes has the same numerical value as that of the electron, and since one hole is created for every electron that is promoted to the conduction band, ne = nh , such that Eq. (6.20) simplifies to σ = ne qe (µe + µh ) (6.21) The mobilities of holes are always less than those of electrons; that is µh < µe . In silicon and germanium, the ratio µe /µh is approximately three and two, respectively (see Table 6.2). Since the mobilities change only slightly as compared to the change of the charge carrier densities with temperature, the temperature variation of conductivity for an intrinsic semiconductor is similar to that of charge carrier density. It can be shown from quantum mechanics that the effective density of states for the conduction and valence bands, Nc and Nv , respectively, are 2πm∗e kB T Nc = 2 h2 

Table 6.2

3/2

(6.22)

Some Room-Temperature Electronic Properties of Group IV Elements

Element C (diamond) Silicon Germanium Gray tin White tin Lead

Eg (eV) ∼7 1.21 0.785 0.09 0 0

σ (ohm−1 · m−1 )

µe (m2 /V · s)

µh (m2 /V · s)

10−14 4.3 × 10−4 2.2 3 × 105 107 5 × 106

0.18 0.14 0.39 0.25 — —

0.14 0.05 0.19 0.24 — —

Source: K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

and 2πm∗h kB T Nv = 2 h2 

3/2

553

(6.23)

where m∗e and m∗h are the effective mass of electrons and holes, respectively, kB is Boltzmann’s constant, T is absolute temperature, and h is Planck’s constant. If we multiply these densities of state by the distribution of Eq. (6.11), in which we replace the energy, E, with the respective energy of the conduction and valence bands (Ev is lower in energy than the Fermi energy, Ef , and Ec is higher), we can obtain the number of mobile charge carriers in each band:

and



−(Ec − Ef ) ne = Nc exp kB T



(6.24)





(6.25)

−(Ef − Ev ) nh = Nv exp kB T

The product of the mobile positive and negative charge carriers is then obtained by multiplication of Eqs. (6.24) and (6.25), and recalling that ne = nh we obtain nh · ne =

n2e



−(Ec − Ev ) = Nc Nv exp kB T



(6.26)

and since Ec − Ev = Eg , the band gap energy is 1/2

ne = (Nv Nc )



−Eg exp 2kB T



(6.27)

Equation (6.27) illustrates that the product of charge carriers, ne nh , is independent of the position of the Fermi level and depends on the energy gap. It is a further result of this relationship that Ec + Ev (6.28) Ef = 2 Thus, the Fermi level is halfway between the conduction and valence bands. The term (Nv Nc )1/2 in Eq. (6.27) depends on the band structure of the semiconductor and is usually a constant for a specific material, outside of its temperature dependence. Thus, we can remove the temperature dependence of Nv and Nc from Eqs. (6.22) and (6.23) and simplify Eq. (6.27) to ne = CT 3/2 exp



−Eg 2kB T



(6.29)

where C is a constant. Substitution of Eq. (6.29) into Eq. (6.21) gives σ = CT 3/2 qe (µe + µh ) exp



−Eg 2kB T



(6.30)

554

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Neglecting the variation of the T 3/2 term, which is negligible compared to the variation with temperature in the exponential term, and recalling that the mobilities are less sensitive to temperature than are the charge carrier densities, Eq. (6.30) can be rewritten as   −Eg (6.31) σ = σ0 exp 2kB T where σ0 is the overall constant. Equation (6.31) indicates that the conductivity of intrinsic semiconductors drops nearly exponentially with increasing temperature. At still higher temperatures, the concentration of thermally excited electrons in the conduction band may become so high that the semiconductor behaves more like a metal. 6.1.1.5 Extrinsic Semiconduction. The charge carrier density can also be increased through impurities of either higher or lower valence. For example, if pentavalent substitutional atoms such as P, As, or Sb are placed into a covalently bonded tetravalent material such as Si or Ge, in a process known as doping, only four of their five valence electrons are required to participate in covalent bonding. Since the fifth electron remains weakly bound to the impurity or donor atom, it is not entirely free in the crystal. Nevertheless, the binding energy, which is of the order 0.01 eV, is much less than that of a covalently bonded electron. This “extra” electron can be easily detached from the impurity or donor atom. The energy state of this electron is indicated by Ed , since it is the donor level, and the energy required to excite it to the conduction band can be represented by the band model shown in Figure 6.13. Notice that the energy gap between the donor and conducting bands is much smaller than that between the valence and conduction bands, or the normal band gap, Eg . Thermal agitation, even at room temperature, is sufficient to transfer this electron to the conduction band, leaving behind a positively charged hole in the donor

Conducting band

Electrons

Ec +

+

+

+

+

+

+

+

+

+

+

Ed Ef

Eg

Ef1

Ev Valence band

Figure 6.13 Schematic illustration of energy bands in an n-type, extrinsic semiconductor. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

555

band. The conductivity in this case is due to motion of electrons in the conduction band, and this type of material is called an n-type semiconductor. This is one type of extrinsic semiconductor, or impurity semiconductor, in which an insulating or intrinsic semiconducting material is made to increase its semiconduction capabilities through the addition of impurity atoms. Notice also, in Figure 6.13, that the Fermi level of the n-type extrinsic semiconductor, Ef , is higher than that of the undoped, intrinsic semiconductor, Ef 1 . From a conductivity standpoint, the other possibility in the addition of impurity atoms is to substitute with an impurity of lower valence, such as a trivalent element in a tetravalent lattice. Examples would be the addition of boron, aluminum, gallium, indium, and thallium to silicon or germanium. In this case, a p-type, extrinsic semiconductor is created. Since the substitutional atoms are deficient in bonding electrons, one of their bonding orbitals will contain a hole that is capable of “accepting” an electron from elsewhere in the crystal. As shown in Figure 6.14, the binding energy is small, and the promotion of an electron from the valence band to the acceptor band, Ea , leaves a hole in the valence band that can act as a charge carrier. In summary, then, for an n-type semiconductor, electrons in the conduction band are charge carriers and holes participate in bonding, whereas for a p-type semiconductor, the electrons participate in bonding, and the holes are the charge carriers. In both types of extrinsic semiconductors, the doping elements are chosen so that both acceptor and donor levels are located closer to the corresponding energy bands and only a small energy gap is involved when exciting electrons. Intrinsic semiconductors are prepared routinely with initial total impurity contents less than one part per million (ppm), along with unwanted donor or acceptor atoms at less than one part in 107 . To such highly purified materials, controlled amounts of substitutional donor or acceptor impurities are intentionally added on the order of 1–1000 ppm to produce extrinsic semiconductors having specific room temperature conductivities. Unlike intrinsic semiconductors, in which the conductivity is dominated by the exponential temperature and band-gap expression of Eq. (6.31), the conductivity of extrinsic semiconductors is governed by competing forces: charge carrier density and charge carrier mobility. At low temperatures, the number of charge carriers initially Conducting band

Ec

Eg

Ef1 Eg Ef −



















+

+

+

+

+

+

+

+

+

+ Holes

Valence band



Ea Ev

Figure 6.14 Schematic illustration of energy bands in a p-type, extrinsic semiconductor. From Z. Jastrzebski; The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

556

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Cooperative Learning Exercise 6.2

The following electrical conductivity characteristics for both the intrinsic and extrinsic forms of a semiconductor have been determined at room temperature. Recall that qe = qh = 1.602 × 10−19 coulombs.

Intrinsic Extrinsic

σ ( · m)−1

ne (m−3 )

nh (m−3 )

2.5 × 10−6

3.0 × 1013

3.0 × 1013

3.6 × 10−5

4.5 × 1014

2.0 × 1012

Person 1: Use the data to write a relationship for the intrinsic conductivity of this substance as a function of the electron and hole mobilities. Person 2: Use the data to write a relationship for the extrinsic conductivity of this substance as a function of the electron and hole mobilities. Assuming that the mobilities are the same in both the intrinsic and extrinsic states, combine your information to solve for the electron and hole mobilities of this substance. Can you tell whether the substance is an n-type of p-type extrinsic semiconductor by looking at the data? Answers: Equation (6.20) can be used for both; 0.52 = µe + µh ; 112.4 = 225µe + µh . Solve simultaneously to get µe = 0.50 m2 /V-s; µh = 0.02 m2 /V-s. This is an n-type semiconductor, since ne ≫ nh .

increases with temperature, as shown in Figure 6.15 for an n-type semiconductor, because thermal activation promotes donor electrons to the conduction band. At intermediate temperatures, most of the donor electrons have been promoted and the number of charge carriers is nearly independent of temperature in what is known as the exhaustion range. At higher temperatures, the number of valence (bonding) electrons excited to the conduction band greatly exceeds the total number of electrons from substitutional donor atoms, and the extrinsic semiconductor behaves like an intrinsic semiconductor (dashed line in Figure 6.15). On the basis of charge carrier density (conduction electrons) alone, then, the conductivity of an extrinsic semiconductor should vary with temperature as shown in Figure 6.16. However, charge carrier mobility has a temperature-dependent effect. As mentioned in the previous section, the temperature dependence of carrier mobility has little effect on intrinsic conductivity. However, when extrinsic conductivity predominates and donor levels are close the conduction band, the carrier mobility can determine the temperature dependence of conductivity. This occurs in the exhaustion range where mobility typically decreases with increasing temperature because of increased phonon scattering. Thus, in this region, characterized by an approximately constant number of charge carriers and a decreasing carrier mobility with temperature, the temperature dependence of conductivity of an extrinsic semiconductor is similar to that of a pure metal (cf. resistivity plot for pure copper in Figure 6.4). In summary, at low temperature the conductivity varies with temperature as the charge carrier concentration, at higher temperatures the charge carrier mobility dominates and the conductivity decreases with

ELECTRICAL PROPERTIES OF MATERIALS

557

10−5

~0.2 ppm As Exhaustion range

~2 × 10−3 ppm As 10−9 Exhaustion range

Intrins

ic Ge

ne (per Ge atom)

10−7

10−11

10−13 101

102 T (K)

103

Figure 6.15 Temperature dependence of charge carrier concentration of n-type extrinsic germanium with two different As impurity levels. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

temperature, and at still higher temperatures the extrinsic semiconductor behaves as an intrinsic semiconductor. This behavior is summarized in Figure 6.16. 6.1.1.6 Semiconductor Junctions and Devices∗ . Semiconductor crystals can be made in such a way that an n-type region can be made adjacent to a p-type region. This is normally accomplished by selective doping during the growth of a single crystal. The boundary between such regions within a single crystal is called the pn junction, and it has important applications to the development of transistors, rectifiers, and solar energy cells. Let us examine some of these devices. The rectifier, or diode, is an electronic device that allows current to flow in only one direction. There is low resistance to current flow in one direction, called the forward bias, and a high resistance to current flow in the opposite direction, known as the reverse bias. The operation of a pn rectifying junction is shown in Figure 6.17. If initially there is no electric field across the junction, no net current flows across the junction under thermal equilibrium conditions (Figure 6.17a). Holes are the dominant carriers on the p-side, and electrons predominate on the n-side. This is a dynamic equilibrium: Holes and conduction electrons are being formed due to thermal agitation. When a hole and an electron meet at the interface, they recombine with the simultaneous emission of radiation photons. This causes a small flow of holes from the p-region

558

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

10−7 103

10−8 Conductivity (s)

10−9

s (Ω−1 m−1)

ne (per Ge atom)

102

Conduction electrons (ne ) 101

10−10 100

10−11 101

102 T (K)

103

Figure 6.16 Temperature dependence of charge carrier density and conductivity of extrinsic semiconductor Ge doped with 2 ppb As. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc. Recombination zone Electron flow − − + − + − + + + − − − − + + − − + + − − +

Hole flow + +

p-side + + + +

+ +

+



+ +

− +

n-side − − − − −



Battery −

Electron flow

Hole flow

(a)

(b) +



+

+ + + − + + + + ++



− − − − + − − − − −

Battery (c)

Figure 6.17 Schematic illustration of electron and hole distribution for a pn-rectifying junction for (a) no applied electrical potential, (b) forward bias, and (c) reverse bias. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 631. Copyright  2000 by John Wiley & Sons, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

559

Current, I

to the n-region and a flow of electrons from the n-region to the p-region. New holes and electrons continue to form to balance this flow, and an equilibrium is established with no external current flowing through the system. If the n-region is connected to the negative terminal and the p-region to the positive terminal of a battery (Figure 6.17b), holes will move from left to right while electrons will move from right to left. These meet each other and recombine as before in the recombination zone. This produces an abundant number of current carriers flowing across the junction; hence the resistance in the forward direction is low. Only a small voltage is necessary to maintain a large current flow. For reverse bias (Figure 6.17c), holes are attracted to the negative terminal, and electrons are attracted to the positive terminal. Thus, electrons and holes move away from each other and the junction region will be relatively free from charge carriers, in which case it is called the barrier region. Recombination will not occur to any appreciable extent, so that the junction is now highly insulative. There may be, however, a small flow of current as the result of the thermal generation of holes and electrons in the semiconductor, termed leakage current. The current–voltage behavior of the pn junction under both forward and reverse biases is shown in Figure 6.18. If the voltage is allowed to vary sinusoidally (Figure 6.19a), as in an alternating current, the current also varies, but the maximum current for the reverse bias, IR , is extremely small compared to that for the forward bias, IF . The pn junction can be fabricated so that the leakage current in the reverse bias is extremely small, and potentials of up to several hundred volts can be applied before breakdown will occur. Breakdown occurs at high reverse voltages (over several hundred volts), when large numbers of charge carriers are generated, giving rise to an abrupt increase in current. Atoms may actually be ionized during breakdown due to the acceleration of electrons and holes in a phenomenon known as the avalanche effect. Some devices, called Zener diodes, are designed to utilize breakdown behavior

+

IF

Forward bias

Breakdown −

−V0 Reverse bias

0

IR

+V0

+ Voltage,V



Figure 6.18 The current–voltage characteristics of a pn-junction. for forward and reverse biases. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 632. Copyright  2000 by John Wiley & Sons, Inc.

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Voltage,V Reverse Forward

560

+V0 0 −V0 Time

Current, I Reverse Forward

(a)

IF

0 IR

Time (b)

Figure 6.19 (a) Voltage–time and (b) current–time for a pn-rectifying junction. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 632. Copyright  2000 by John Wiley & Sons, Inc.

as voltage regulators. The voltage breakdown can be varied by controlling the dopant level; generally, the higher the impurity concentration, the lower the breakdown voltage. In silicon semiconductors, for example, the breakdown voltage can be varied from about 10 V to about 1000 V by regulating the impurity concentration. Another type of semiconductor junction device is a transistor, which is capable of two primary types of functions. The first is to amplify an electrical signal. The second is to serve as a switching device for the processing and storage of information. We will briefly describe the junction transistor here, and we will defer the description of another type of transistor called a MOSFET to Section 6.1.2.6. The junction transistor is composed of two pn junctions arranged back to back in either the p-n-p or n-p-n configurations. The latter configuration is shown in Figure 6.20. It is composed of a thin n-type base region, sandwiched between a p-type emitter and a p-type collector. The first pn junction is forward-biased, and the second np junction is reverse-biased. Since the emitter is p-type and Junction 1 is forward biased, a large number of holes enter the n-type base region. These holes combine with electrons in the base. However, in a properly prepared junction transistor, the base is thin enough to allow most of the holes to pass through the base without recombination and into the p-type collector. A small increase in input voltage within the emitterbase circuit produces a large increase in current across Junction 2 due to the passage of holes. The large increase in collector current also results in a large increase in voltage across the load resistor. Thus, a voltage signal that passes through a junction transistor experiences amplification, as illustrated by the input and output voltage plots in Figure 6.20. A similar effect is observed in an n-p-n junction transistor, except that electrons instead of holes are injected across the base to the collector.

ELECTRICAL PROPERTIES OF MATERIALS

Emitter p

Junction 2 Base n

Collector p



+ − Load Input voltage

Output voltage

+

Reverse-biasing voltage

Forward-biasing voltage

Junction 1

561

10 Input voltage (mV)

Output voltage (mV)

0.1

Time

Time

Figure 6.20 Schematic diagram of a pnp-junction transistor and associated input and output voltage characteristics. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 633. Copyright  2000 by John Wiley & Sons, Inc.

In addition to their ability to amplify electrical signals, transistors and diodes may act as switching and information storage devices. Computer information is stored in terms of binary code—for example, zeros and ones. Transistors and diodes can operate as switches that have two states: on and off, or conducting and nonconducting. Thus, a conducting circuit can represent a zero while a nonconducting circuit can represent a one. In this way, information and logical operations can be stored and retrieved from a series of semiconducting circuits known as integrated circuits. Inexpensive integrated circuits are mass produced by forming layers of high-purity silicon in a precisely detailed pattern and introducing specific dopants into specific areas by diffusion or ion implantation. The details of some of these processes will be described in Chapter 7. 6.1.2

Electrical Properties of Ceramics and Glasses

Many of the fundamental relationships and concepts governing the electrical properties of materials have been introduced in the previous section. In this section, we elaborate upon those topics that are more prevalent or technologically relevant in ceramics and glasses than in metals, such as electrical insulation and superconductivity, and introduce some topics that were omitted in Section 6.1.1, such as dielectric properties. 6.1.2.1 Electrical Resistance. At room temperature, many ceramics (and polymers, for that matter) have electrical resistivities that are approximately 20 orders of magnitude higher than those of metals (see Figure 6.21). As a result, they are often used as electrical insulators. With rising temperature, however, insulating materials experience an increase in electrical conductivity, which may ultimately be greater than that for semiconductors. In fact, some oxide ceramics such as ReO3 can achieve metallic levels of electrical conductivity at high temperatures. In all cases, the electrical conductivity can be interpreted in terms of carrier concentrations and carrier mobilities as previously described, though we shall see in Section 6.1.2.3 that the carriers may be different for ceramics and glasses than for metals.

562

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

1016 Sintered alumina

1014

Resistivity (Ω cm)

Low-loss steatite 1012 Zircon porcelain

1010

Steatite

108

106 High-voltage porcelain 104

0

200

400

600

800

1000

1200

Temperature (°C)

Figure 6.21 Temperature dependence of resistivity for some typical oxide ceramics. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Despite the potential for high electrical conductivities, we will concentrate on the more common high-resistivity aspects of ceramics in this section and will describe some of the factors that affect the use of ceramics and glasses as electrical insulators. For ceramics and glasses having electrical resistivities above about 105  · cm, the concentration–mobility product of the charge carriers is small; as a result, minor variations in composition, impurities, heat treatment, stoichiometry, porosity, oxygen partial pressure, and other variables can have a significant effect on the resistivity. The effect of porosity is similar to that described for thermal conductivity in Section 4.2. As porosity increases, the electrical resistivity increases proportionally for small values of porosity. At higher porosities, the effect is more substantial. The effect of grain boundaries in polycrystalline materials is related to the mean free path of the ions, as described in Section 6.1.1.1. Except for very thin films or extremely fine-grained materials (less than 0.1 µm), the effects of grain boundary scattering are small compared to lattice scattering. Consequently, the grain size in uniform-composition materials has little effect on resistivity. However, substantial effects can result from impurity concentrations at grain boundaries. Particularly for oxide materials, there is a tendency to form a glassy silicate phase at the boundary between particles. Ceramic compositions with the highest glassy phase have the lower resistivities at moderate temperatures. In glasses, the alkali content should be kept to a minimum when high resistivities are required. The reasons for this are related to the electrical conduction via ions, which will be described in Section 6.1.2.3. For now, we maintain our emphasis on the electrical insulating properties of ceramics and glasses in order to describe a unique application known as dielectrics. 6.1.2.2 Dielectrics, Ferroelectricity, and Piezoelectricity. A dielectric is a material separating two charged bodies, as illustrated in Figure 6.22b. For a substance

ELECTRICAL PROPERTIES OF MATERIALS

D0 =

563

0

V

l

= V l

Vacuum

(a)

D0 =

0

+P

V

Dielectric

− +

P









+

+

+

+

(b)

Figure 6.22 A parallel-plate capacitor (a) in vacuum and (b) with a dielectric material between the plates. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 640. Copyright  2000 by John Wiley & Sons, Inc.

to be a good dielectric, it must be an electrical insulator. As a result, any electrical insulator is also called a dielectric. The interaction of a dielectric with an electric field is conveniently described in terms of a parallel plate capacitor. When a voltage is applied between two parallel plates in a vacuum (Figure 6.22a), one plate becomes positively charged and the other becomes negatively charged. If an electrical conductor were to be placed in contact with the plates, current would flow. In a vacuum, however, no current flows, and the electric field strength is a function of the applied voltage, V , and the separation distance between the plates, l, as given earlier by Eq. (6.4): E = V/l

(6.4)

The magnitude of the charge per unit area on either plate, called the electric displacement or flux density, D0 , is directly proportional to the field: D 0 = ε0 E

(6.32)

564

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

where ε0 is a constant called electric permittivity and has a value of 8.854 × 10−12 C/V · m (or farad/m, F/m, where a farad equals a coulomb/volt). The electric displacement therefore has units of C/m2 . Notice the similarity in form of Eq. (6.32) to Eq. (6.2) for the magnetic flux in a magnetic field. As before, both D0 and E are vector quantities, but for our purposes a description of their magnitudes is sufficient, so the boldface will be dropped. Since D0 is the surface charge density on the plate in vacuum, it can be related to the capacitance of a parallel plate capacitor in vacuum, C0 , which is defined as C0 = q/V

(6.33)

where q is the magnitude of charge on each plate and V is the applied voltage. With A the plate area and l the separation between plates, along with q = D0 A, Eq. (6.33) becomes A (6.34) C0 = ε0 l when a vacuum exists between the plates. If we now place a dielectric (insulator) material between the plates, as in Figure 6.22b, a displacement of charge within the material is created through a progressive orientation of permanent or induced dipoles. Recall that a dipole is simply any type of charge separation, as first described in Section 1.0.3.2 and illustrated in Figure 6.23, and that it may be permanent or temporarily induced. The magnitude of the electric dipole, pe , is simply proportional to the charge, q, and the separation distance, a, as pe = qa. The interaction between permanent or induced electric dipoles with an applied electric field is called polarization, which is the induced dipole moment per unit volume. Polarization causes positive charge to accumulate on the bottom surface next to the

Center of positive charge

+ − Dipole moment

− + Center of negative charge

Figure 6.23

Schematic illustration of an electric dipole.

ELECTRICAL PROPERTIES OF MATERIALS

565

negatively charged plate and causes negative charge to accumulate toward the positively charged plate on the top. This tends to decrease the effective surface charge density on either plate. The expected decrease in effective charge corresponds to the polarization, P , of the material, which equals the induced dipole moment per unit volume of polarizable material. The electric field is now found from Eq. (6.32) by accounting for the reduction in displacement due to polarization: E = (D − P )/ε0

(6.35)

The displacement can now be found by rearranging Eq. (6.2) to D = ε0 E + P

(6.36)

D = εE

(6.37)

Equation (6.36) is often written

where ε is called the permittivity. The ratio of permittivity with the dielectric to the permittivity in vacuum, ε/ε0 , is called the relative permittivity, εr , or dielectric constant. The dielectric constant is a material property. Some values of dielectric constants for common ceramic and glass insulators are given in Table 6.3. Since a polarizable material causes an increase in charge per unit area on the plates of a capacitor, the capacitance also increases, and it can be shown that the dielectric constant is related to the capacitance and displacement in vacuum and with the dielectric material as follows: εr =

D C ε = = ε0 D0 C0

(6.38)

Cooperative Learning Exercise 6.3

A voltage of 10 V is applied across a parallel-plate capacitor with a plate separation of 2 × 10−3 m and plate area of 6.45 × 10−4 m2 , both in vacuum and then with a dielectric material placed between the plates. The polarization, P , due to the presence of the dielectric material is 2.22 × 10−7 C/m2 . Person 1: Calculate the displacement in vacuum, D0 , and the capacitance in vacuum, C0 . Person 2: Calculate the displacement, D, due to the presence of the dielectric. Combine your information to calculate the dielectric constant and the capacitance in the presence of the dielectric. Answers: Eq. (6.32) D0 = 8.854 × 10−12 (10/0.002) = 4.427 × 10−8 C/m2 , Eq. (6.34) C0 = 8.854 × 10−12 (6.45 × 10−4 /0.002) = 2.85 × 10−12 F; Eq. (6.36) D = 8.854 × 10−12 (10/0.002) + 2.22 × 10−7 = 2.66 × 10−7 C/m2 . εr = D/D0 = 6.0; C = εr C0 = 1.71 × 10−11 F.

566

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Table 6.3

Dielectric Constants of Selected Ceramics and Glasses at 25◦ C and 10◦ Hz

Glasses Silica glass Vycor glasses Pyrex glasses Soda–lime–silica glass High-lead glass

εr 3.8 3.8–3.9 4.0–6.0 6.9 19.0

Inorganic Crystalline Materials Barium oxide Mica Potassium chloride Potassium bromide Cordierite ceramics (based on 2MgO · 2Al2 O3 · 3SiO2 ) Diamond Potassium iodide Forsterite (Mg2 SiO4 ) Mullite (3Al2 O3 · 2SiO2 ) Lithium fluoride Magnesium oxide

εr 3.4 3.6 4.75 4.9 4.5–5.4

5.5 5.6 6.22 6.6 9.0 9.65

Source: K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc.

We can further describe the polarization, P , according to the different types of dipoles that either already exist or are induced in the dielectric material. The polarization of a dielectric material may be caused by four major types of polarization: electronic polarization, ionic (atomic) polarization, orientation polarization, and spacecharge (interfacial) polarization. Each type of polarization is shown schematically in Figure 6.24 and will be described in succession. In these descriptions, it will be useful to introduce a new term called the polarizability, α, which is simply a measure of the ability of a material to undergo the specific type of polarization. Electronic polarization arises because the center of the electron cloud around a nucleus is displaced under an applied electric field, as shown in Figure 6.24. The resulting polarization, or dipole per unit volume, is Pe = N αe Eloc

(6.39)

where N is the number of atoms or molecules, αe is the electronic polarizability, and Eloc is the local electric field that an atom or molecule experiences. The local electric field is greater than the average field, E , because of the polarization of other surrounding atoms. Ionic polarization occurs in ionic materials because an applied field acts to displace cations in the direction of the applied field while displacing anions in a direction opposite to the applied field. This gives rise to a net dipole moment per formula unit. For an ionic solid, the atomic polarization is given by Pi = N αi Eloc

(6.40)

where N is the number of formula units per unit volume, and αi is the ionic polarizability. Both ionic and electronic polarization contribute to the formation of induced dipoles.

ELECTRICAL PROPERTIES OF MATERIALS

No field

567

Field applied Electronic polarization

+

+

Ionic polarization

+



+



Orientation polarization

Space charge polarization +



+



+

+

+









+



+



+



+

+



+



+



+

+

+

+







+



+



+

+







Figure 6.24 Schematic illustration of different types of polarization. From W. D. Kingery, H. K. Bowen and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Orientation polarization can occur in materials composed of molecules that have permanent electric dipole moments. The permanent dipoles tend to become aligned with the applied electric field, but entropy and thermal effects tend to counter this alignment. Thus, orientation polarization is highly temperature-dependent, unlike the forms of induced polarization which are nearly temperature-independent. In electric fields of moderate intensity, the orientation polarization is proportional to the local electric field, as for the other forms of polarization Po = N αo Eloc

(6.41)

but in strong electric fields this proportionality is not maintained since saturation must occur when all the permanent dipoles are aligned. Orientation polarization occurs mostly in gases and liquids where molecules are free to rotate. In some solids the rotation of polar molecules may be restricted by lattice forces, which can reduce orientation polarization. For example, the dielectric constant of water is about 82, while that of ice is about 10. The orientation polarizability, αo , can be further elaborated

568

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

upon, since the permanent electric dipole moment per molecule, pe , is a measurable quantity. The two quantities are related by a relationship similar to the Curie Law, which we shall elaborate upon in Section 6.2: αo =

pe 2 3kB T

(6.42)

where kB is Boltzmann’s constant and T is absolute temperature. A list of permanent dipole moments for selected polar molecules is given in Table 6.4. The final type of polarization is space-charge polarization, sometimes called interfacial polarization, and results from the accumulation of charge at structural interfaces in heterogeneous materials. Such polarization occurs when one of the phases has a much higher resistivity than the other, and it is found in a variety of ceramic materials, especially at elevated temperatures. The space-charge polarization, Psc , has a corresponding space-charge polarizability, αsc . The two are related via a relationship of the form for the other types of polarization. The total polarization, then, is the sum of all the contributions from the different types of polarization: P = Pe + Pi + Po + Psc (6.43) Table 6.4 Permanent Dipole Moments of Some Polar Molecules in the Gaseous Phase Molecule

1030 × Dipole Moment(C · m)

AgCl LiCl NaCl KCl CsCl LiF NaF KF RbF CsF HF HCl HBr HI HNO3 H2 O NH3 CH3 Cl CH2 Cl2 CHCl3 C2 H5 OH C6 H5 NO2

19.1 23.78 30.0 34.26 34.76 21.1 27.2 28.7 27.5 26.3 6.07 3.60 2.74 1.47 7.24 6.17 4.90 6.24 5.34 3.37 5.64 14.1

Source: K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

569

With appropriate substitution of the expressions for the various types of polarization, it is easy to show that the total polarizability, α, is also the sum of the various types of polarizability: (6.44) α = αe + αi + αo + αsc The total polarization, P , and hence the dielectric constant, εr , subjected to an alternating electric field depend upon the ease with which the permanent or induced dipoles can reverse their alignment with each reversal of the applied field. The time required for dipole reversal is called the relaxation time, and its reciprocal is called the relaxation frequency. When the time duration per cycle of the applied field is much less than the relaxation time of a particular polarization process, the dipoles cannot change their orientation rapidly enough to remain oriented in the applied field, and they “freeze” their reorientation process. This particular polarization process will not contribute to the total polarization. The order in which the polarization processes cease to contribute to the total polarization is related to their relative size scales. Obviously, electronic polarization is the smallest since it deals with electron density. Ionic polarization is next in size scale, since it deals with the separation of ions, followed by orientation polarization, which is on the order of molecular dimensions and can sometimes be large. Finally, space-charge polarization, if it exists at all, is on the scale of interfacial dimensions. The larger the size scale, the more difficult it is for that type of polarization to respond to an alternating electric field. Thus, we would expect space charge to “freeze” up first with increasing electric field, followed by orientation, ionic, and electronic polarization, respectively. This process is shown in a plot of polarizability (which recall is proportional to polarization) versus the logarithm of the alternating electric field frequency in Figure 6.25a. As the frequency of the applied field approaches the relaxation frequency, the polarization response increasingly lags behind the applied field. The reorientation of each type of dipole is opposed by internal friction, which leads to heating in the sample and power loss. The power loss depends on the degree to which the polarization lags behind the electric field. This lag can be quantified by measurement of the displacement–field relationship in a sinusoidally varying field. It is found that if the alternating field is given by E = E0 sin(ωt) (6.45) then the polarizability, P , is given by P = P0 sin(ωt − δ)

(6.46)

where δ is the phase angle that the polarization lags behind the field, and ω is the angular frequency of the field. It can be shown that the power loss is related to the phase angle and the applied field: Power loss = 2π E 2 ω ε0 tan δ

(6.47)

where tan δ is called the loss tangent. For small phase angles, tan δ ≈ δ. As shown in Figure 6.25b, the power loss for a particular polarization type increases as that process becomes slower to respond to frequency changes, reaches a maximum as the electric field frequency reaches the relaxation frequency, and decreases after that polarization type has reached its relaxation frequency.

570

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Total polarizability

Electronic Ionic Orientation Space charge

Sorce charge

Orientation Ionic Electronic

Power loss

(a)

Power Audio Radio Electrical frequencies

Infrared

Visible

Optical frequencies

log w (b)

Figure 6.25 Frequency dependence of (a) total polarizability and (b) power loss. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

The loss tangent is an important quantity. The product of the loss tangent and the dielectric constant, εr tan δ, is called the loss factor and is the primary criterion for judging the usefulness of a dielectric as an insulator material. For this purpose, it is desirable to have a low dielectric constant and a small loss angle. For applications in which it is desirable to obtain a high capacitance in the smallest physical space, a high dielectric constant must be used, and a low value of the loss angle is needed. At high enough frequencies, the dielectric will experience electrical breakdown and the insulator will become a conductor. As in a semiconductor, breakdown is initiated by the field-induced excitation of a number of electrons to the conduction band with sufficient kinetic energy to knock other electrons into the conduction band. The result is the avalanche effect described earlier. The magnitude of the electric field required to cause dielectric breakdown is called the dielectric strength, or breakdown strength. The dielectric strength for oxide ceramic insulators lies in the range 1–20 kV/mm, with the value of 9 kV/mm for Al2 O3 being typical.

ELECTRICAL PROPERTIES OF MATERIALS

571

The temperature dependence of the dielectric constant is highly dependent upon molecular structure, as well as the predominant polarization mechanism. For large molecules, orientation polarization is difficult below the material melting point, whereas orientation polarization can persist below the melting point in smaller molecules. Some substances undergo structural transformations in the solid state from higher- to lowerordered structures which can restrict molecular rotation. The effects of temperature and frequency on the dielectric constant are not entirely independent. For example, the frequency effect on electronic and ionic polarization is negligible at frequencies up to about 1010 Hz, which is the limit of normal use. Similarly, the effect of temperature on electronic and ionic polarization is small. At higher temperatures, however, there is an increasing contribution resulting from ion mobility and crystal imperfection mobility. The combined effect is to give a sharp rise in the apparent dielectric constant at low frequencies with increasing temperature, as illustrated for alumina porcelain, whose primary component is Al2 O3 , in Figure 6.26. The loss tangent is also shown 15 Relative dielectric constant

102

103 104

10 106,1010 5

0

102

1

Tan d

103

104

106

0.01

1010 0.001

0.0001

0

100

200

300

400

500

Temperature (°C)

Figure 6.26 Frequency and temperature dependence of dielectric constant and tan δ for alumina porcelain. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

572

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

in Figure 6.26 for reference. Alumina is an important component of most ceramic dielectrics. The other base components are silica, SiO2 , and magnesia, MgO. Two phenomena related to the electric dipoles in a material are ferroelectricity and piezoelectricity. Ferroelectricity is defined as the spontaneous alignment of electric dipoles by their mutual interaction in the absence of an applied electric field. The source of ferroelectricity arises from the fact that the local field increases in proportion to the polarization. Thus, ferroelectric materials must possess permanent dipoles. One of the most important ferroelectric materials is barium titanate, BaTiO3 . The spontaneous polarization of BaTiO3 is related directly to the positions of the ions within the tetragonal unit cell, as shown in Figure 6.27a. Divalent barium ions occupy each corner position, with the tetravalent titanium ion near, but just above, the center of the cell. The oxygen ions are located below the center of the (001) plane and just below the centers of the (100) and (010) planes, as illustrated in Figure 6.27b. Consequently, per unit cell, the centers of positive and negative charge do not coincide, and there is spontaneous polarization. At a certain temperature called the ferroelectric Curie temperature (120◦ C for barium titanate), barium titanate becomes cubic with Ti4+ ions located at the center of the unit cell and O2− ions occupying face-centered positions, and it loses its spontaneous polarization. Ferroelectric materials have extremely high dielectric constants at relatively low applied field frequencies. Consequently, capacitors made from these materials can be significantly smaller than capacitors made from other dielectric materials. Closely related to ferroelectricity is piezoelectricity in which polarization is induced and an electric field is established across a specimen by the application of external force (see Figure 6.28a,b). Reversing the direction of the external force, as from tension to compression, reverses the direction of the field. Alternatively, the application of an external electric field alters the net dipole length and causes a dimensional change, as in Figure 6.28c. Piezoelectric materials can be used as transducers—devices that

Ba

O 0.403 nm

0.009 nm O2− 0.006 nm

Ti

Ba2+ 0.006 nm 0.398 nm

Ti4+

0.398 nm (a)

(b)

Figure 6.27 View of (a) the tetragonal unit cell of BaTiO3 and (b) the ionic positions projected in a [100] direction onto a (100) face. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

573

s ++ + + + + + + + + + + + + + + + ++++ ++ + + + + + + + + +

+ + + + + +

+++++++++++++++++++







+

+

+







+

+

+

−−−−−−−−−−−−−−−−−−−

V0

− +

− +

− +

V < V0

− − − + + + − − − − − −







+

+

+







+

+

+

− − − − − − − − − − − − − − − − − − − − − −

s (a)

(b)

(c)

Figure 6.28 Schematic illustration of the piezoelectric effect that occurs in (a) an unstressed and (b) stressed piezoelectric material. Mechanical deformation can also occur when (c) a voltage is applied to a piezoelectric material. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

convert mechanical stress into electrical energy and vice versa. Such devices are used in microphones and mechanical strain gauges. Barium titanate is also a piezoelectric material. Before proceeding, it may be useful to summarize some of the myriad of terms and symbols that have been used to this point. Such a summary is presented in Table 6.5. With the exception of a few more terms dealing with superconductivity, we have introduced all the variables we will need to describe the electrical properties of materials. 6.1.2.3 Ion-Conducting Ceramics and Glasses. In addition to the conduction of charge via electrons, charge can be conducted via ions. Ions are present in most crystalline ceramic materials such as oxides and halides. This process is termed ionic conduction and may occur either in conjunction with or separately from electronic conduction. As a result, we must expand our definition of conductivity to include both types of conduction: σ = σelectronic + σionic (6.48)

In a manner analogous to conduction via electrons and holes, the ionic conductivity can be related to the charge carrier density, ni , charge carrier mobility, µi , and carrier charge, qe , as originally given in Eq. (6.9) for electronic conduction, except that the charge carriers are now ions, which have multiple charges, or valences, zi , so that Eq. (6.9) must be modified slightly: σionic = f ni (zi qe )µi

(6.49)

An additional factor, f , has been included to account for the number of equivalent sites to which an ion may transfer in a particular crystal structure; for example f = 4 for an ion vacancy in the cubic crystal structure. For an ion to move through the lattice under an electric field driving force, it must have sufficient thermal energy to pass over an energy barrier, which in this case is

574

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Table 6.5

Electrical Symbols, Terms, and SI Units SI Units

Quantity

Symbol

Primary

Derived

q

coulomb, C

(none)

Electric potential

V

2

Electric current

I

Electric field strength

ε

Resistance

R

Resistivity

ρ

kg · m3 /s · C2

·m

Conductance

G

s · C2 / kg · m2

−1

Electric charge

Conductivity

σ

Electrical capacitance

C

2

kg · m /s · C

volt, V

C/s

ampere, amp, A

kg · m/s2 · C

V/m

kg · m2 /s · C2

2

s · C / kg · m

ohm, 

3

2

( · m)−1

s2 · C / kg · m2

farad, F

2

2

3

s · C / kg · m

Permittivity

ε

Dielectric Constant (Relative Permittivity)

εr

(dimensionless)

(dimensionless)

Dielectric displacement

D

C/m2

Electric polarization

P

F · V/m2

C/m2 2

F/m

2

F · V/m2

debeye, D

Polarizability

α

s · C / kg

Permanent Dipole Moment

p

C·m

F · m2

Dielectric loss angle

δ

(none)

(none)

Charge carrier mobility

µ

Charge carrier density

n

s · C/kg

m2 /V · s

Magnetic field strength

H

C/m · s

A/m

m−3

(none)

Items in boldface are material properties.

an intermediate position between lattice sites. This activation energy barrier to motion has already been described in Section 4.3.2.1 for diffusion in crystalline solids, so it will not be described here. It makes sense, though, that the ion mobility, µi , should be related to the ion diffusivity (diffusion coefficient), Di , and can be derived from Eq. (6.14) presented earlier: zi qe Di (6.50) µi = kB T where zi is the valence of the ion, qe is the electronic charge, and kB and T are Boltzmann’s constant and absolute temperature, as before. Substitution of Eq. (6.50) into Eq. (6.49) gives a relationship for the ionic conductivity: σionic =

f ni (zi qe )2 Di kB T

(6.51)

Typically, there is a predominate charge carrier in ionic solids. For example, in sodium chloride, the mobility of sodium ions is much larger than the mobility of chloride ions. This phenomenon can be temperature-dependent, however. As shown in Table 6.6, the fraction of total conductivity attributable to positive ions (K+ ) in

575

ELECTRICAL PROPERTIES OF MATERIALS

Table 6.6 Ionic and Electronic Contributions to Total Electrical Conductivity for Some Selected Compounds and Glasses

Compound NaCl KCl KCl + 0.02% CaCl2

Temperature (◦ C)

Cationic Conductivity + /σtotal ) (σionic

Anionic Conductivity − /σtotal ) (σionic

Electronic Conductivity (σe,h /σtotal )

400

1.00





600

0.95

0.05



435

0.96

0.04



600

0.88

0.12



430

0.99

0.01



600

0.99

0.01



AgCl

20–350

1.00





AgBr

20–300

1.00





BaF2

500



1.00



PbF2

200



1.00



CuCl

20





1.00

366

1.00





>700



1.00

10−4

Na2 O · 11Al2 O3

77 K, are of great technological interest. Another striking difference the HTSC compounds exhibited was the anomalous behavior when placed in a magnetic field. Recall from Section 6.1.1.3 that there is also a critical magnetic field, Hc , above which the superconductor loses its superconductivity, and that this critical field is a function of temperature. In the metal and compound-based superconductors of Section 6.1.1.3, the transition from superconducting to nonsuperconducting at a specified temperature is a relatively sharp one as the critical field is reached. This relationship was illustrated in Figure 6.10 and Eq. (6.19). When compounds belonging to the new class of HTSC are placed in a magnetic field, however, the temperature region for the superconducting–nonsuperconducting transition broadens as the field is increased, as illustrated for YBa2 Cu3 O7 in Figure 6.31. At low fields the resistivity drops sharply over a very narrow temperature range, whereas at high fields the transition range is quite large—nearly 15 K. Compare this resistivity–temperature behavior with that of Pb in Figure 6.6. Thus, two classes of superconductors are defined: Type I, also called “hard” superconductors, which exhibit a sharp transition; and Type II, also known as “soft” superconductors, which have a range of transition values. In Type II superconductors, the magnetic field starts penetrating into the material at a lower critical field, Hc1 . Penetration increases until at the upper critical field, Hc2 , the material is fully penetrated and the normal state is restored. In addition to a critical temperature and critical field, all superconductors have a critical current density, Jc , above which they will no longer superconduct. This limitation has important consequences. A logical application of superconductors is as current-carrying media. However, there is a limit, often a low one, to how much current they can carry before losing their superconducting capabilities. The relationship between Jc , Hc , and Tc for a Type II superconductor is shown in Figure 6.32. Notice that the Hc –Tc portion of this plot has already been presented in Figure 6.10 for a Type I superconductor.

1

Resistivity (m Ω•cm)

Y Ba2Cu3Ox (single crystal) H//c-axis H (kOe) 0.5

0 3 10 20 40 60 90

0 75

80

85

90

95

Temperature (K)

Figure 6.31 Variation of superconducting transition range in a magnetic field. Reprinted, by permission, from T. P. Sheahan, Introduction to High-Temperature Superconductivity, p. 121. Copyright  1994 by Plenum Press.

ELECTRICAL PROPERTIES OF MATERIALS

579

Current density J

JC (T − 0 K, H = 0)

Hc 2 (T = 0 K, J = 0)

TC (H = 0, J = 0)

Temperature T

Magnetic field H

Figure 6.32 The relationship between temperature, magnetic field and current density in a Type II superconductor. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 699. Copyright  2000 by John Wiley & Sons, Inc.

Table 6.7 Some HTSC Materials Based on the Copper Oxide Lattice Compound YBa2 Cu3 O7

Tc (K) 92

(BiPb)2 Sr2 Ca2 Cu3 Ox

105

TlBa2 Ca2 Cu3 Oy

115

HgBa2 Ca2 Cu3 Oy

135

Most of the oxide Type II HTSC are based upon the copper oxide lattice. Some examples and their transition temperatures are given in Table 6.7. In all cases, the structure is essentially a laminar one, with planes of copper oxide in the center (see Figure 6.33), through which the superconducting current flows. These materials are thus highly anisotropic—very little current can flow perpendicular to the copper oxide planes. As a result, great pains are taken to produce these materials in a highly oriented manner and to use them such that current flows parallel to the copper oxide planes. The role of the elements other than copper and oxygen is secondary. In YBCO, yttrium is only a spacer and a contributor of charge carriers. Nearly any of the rare earth elements (holmium, erbium, dysprosium, etc.) can be substituted for yttrium without significantly changing Tc .

580

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Y

c

Ba b Cu

a

O

Figure 6.33 Unit cell structure of YBa2 Cu3 O7 superconductor. Reprinted, by permission, from Sheahan, T. P., Introduction to High-Temperature Superconductivity, p. 5. Copyright  1993 by Plenum Press.

It is beyond the scope of this text to describe the mechanism for charge transfer in the high-temperature superconductors. While there is still a great deal of discussion on the applicability of BCS theory (cf. Section 6.1.1.3) to HTSC materials, it is safe to say that many of the principles still apply—for example, density of states at the Fermi level. The interested reader should refer to existing literature [3,4] for more information on the structure and theory of copper oxide superconductors. Finally, it would be misleading to suggest that these materials are the technological revolution they were first imagined to be. Like most other oxide ceramics, these materials are very brittle, and attempts to fabricate them in useful forms, such as wire, have proved problematic. The additional difficulty is with current density, as described earlier. For large-scale applications such as particle accelerator magnets, current densities on the order of 1010 A/m2 are required. A high critical magnetic field would also be beneficial for some applications, where magnetic fields of 30 tesla (30,000 gauss) are required. Oxide superconductors meet neither of these requirements. As a result, the metallic superconductors are still the materials of choice for these high-end applications, since they are more easily formed into wires, and have higher Jc and Hc values, even though their Tc is low. Metallic superconductors of the niobium family are particularly useful in these regards and include NbTi, Nb3 Sn, NbZr, Nb3 Al, and Nb3 Ge. 6.1.2.5 Compound Semiconductors. The niobium-based superconducting compounds lead us naturally into another use for intermetallics—namely, semiconductors. This topic, too, was introduced earlier in this chapter (cf. Section 6.1.1.4 and 6.1.1.5), and we shall build upon those principles here to describe the semiconducting properties of compounds, ceramics, and glasses. The classification of intermetallics as ceramics

ELECTRICAL PROPERTIES OF MATERIALS

581

may be confusing at this point, but we do so due to the strong covalent nature of their bonds that are more similar to oxide ceramics than to the metallic bond of elemental semiconductors. It is this bond structure that influences the electrical properties; hence, we discuss them together in this section. Recall that elemental semiconductors such as silicon, germanium, tin, and diamond belong to group IV in the periodic table. They are characterized by four valency electrons, resulting in a tetrahedral arrangement of atoms in a diamond-like structure (cf. Figure 1.40). A large number of compounds possess structural and electrical characteristics that are similar to those of elemental semiconductors, many of which are compounds formed between elements of groups III and V of the periodic table, termed III–V compounds. In these cases, slight deviations from stoichiometry can lead to extrinsic conduction. One of the more successful examples is gallium arsenide, GaAs. Gallium is a group III element and arsenic a group V element, and the resulting average valency in a 50–50 compound is four, just as for a group IV element. However, since the elemental substances have different electronegativities (cf. Table 1.4), the bond will be partly ionic in nature and the band gap will be higher than in the corresponding elemental semiconductor, Ge (see Table 6.8; some values may differ slightly from Table 6.2). These larger band gaps extend the useful range of conduction since intrinsic conduction becomes important only at correspondingly higher temperatures. The general decrease in band gap with increasing atomic number of the II–V compounds is a result of an increasing tendency toward metallic bonding. Some semiconducting compounds can be of the II–VI type, which also has an average valence of four, but these have much more ionic character than III–V compounds. Their band gaps are thus larger, and in some cases they may even be viewed as insulators. For example, ZnS, with a band-gap energy of 3.6 eV, is an insulator, whereas ZnSe has an band gap of 2.8 eV, which is closer to a semiconductor. A wide variety of

Table 6.8 Electronic Properties of Some Elemental and Compound Semiconductors at Room Temperature

Energy

Mobility, µ m2 V ·s

Effective Mass, m∗ /m0

Material

Gap (eV)

Electrons

Holes

Electrons

Holes

Dielectric Constant

Silicon (Si) Ge Sn (gray) GaAs GaSb GaP InSb InAs InP PbS

1.1 0.67 0.08 1.40 0.77 2.24 0.16 0.33 1.29 0.40

0.135 0.390 0.200 0.800 0.400 0.050 7.800 3.300 0.460 0.060

0.048 0.190 0.100 0.025 0.140 0.002 0.075 0.040 0.015 0.020

0.97 1.60

0.5 0.3

11.7 16.3

0.072 0.047

0.5 0.5

0.013 0.02 0.07

0.60 0.41 0.40

12.5 15.0 10.0 17.0 14.5 14.0

Source: Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc.

582

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

compounds between transition metals and nonmetals also exhibit semiconductivity, but these materials are used infrequently because impurity control is much more difficult to achieve. For nonstoichiometric compounds, the general rule is that when there is an excess of cations or a deficiency of anions, the compound is an n-type semiconductor. Conversely, an excess of anions or deficiency of cations creates a p-type semiconductor. There are some compounds that may exhibit either p- or n-type behavior, depending on what kind of ions are in excess. Lead sulfide, PbS, is an example. An excess of Pb2+ ions creates an n-type semiconductor, whereas an excess of S2− ion creates a p-type semiconductor. Similarly, many binary oxide ceramics owe their electronic conductivity to deviations from stoichiometric compositions. For example, Cu2 O is a well-known p-type semiconductor, whereas ZnO with an excess of cations as interstitial atoms is an n-type semiconductor. A partial list of some impurity-controlled compound semiconductors is given in Table 6.9. The conductivity of these materials can be controlled by the number of defects. In a p-type semiconductor such as Cu2 O, in which vacancies are formed in the cation lattice when the oxygen partial pressure is increased, we can develop relationships between conductivity and oxygen partial pressure. The overall reaction for the formation of vacancies and electron holes can be written in Kroger–Vink notation (cf. Section 1.2.6.1) as 1 O (g) OO + 2 V′Cu + 2hž (6.52) 2 2 From this reaction, we can then write a mass-action relation [cf. Eq. (3.1)] that is a function of temperature: [V′ ]2 [hž ]2 (6.53) K(T ) = Cu 1/2 PO2 Table 6.9

Some Impurity Compound Semiconductors n-Type

TiO2

Nb2 O5

CdS

Cs2 Se

BaTiO3

Hg2 S

V2 O5

MoO2

CdSe

BaO

PbCrO4

ZnF2

U3 O8

CdO

SnO2

Ta2 O5

Fe3 O4

ZnO

Ag2 S

Cs2 S

WO3 p-Type

Ag2 O

CoO

Cu2 O

SnS

Cr2 O3

SnO

Cu2 S

Sb2 S3

MnO

NiO

Pr2 O3

CuI

Bi2 Te3

Hg2 O

Amphoteric Al2 O3

SiC

PbTe

Mn3 O4

PbS

UO2

Co3 O4

PbSe

IrO2

MoO2

Ti2 S

ELECTRICAL PROPERTIES OF MATERIALS

583

If the concentration of vacancies is largely determined by reaction with the atmosphere, then the conductivity should be proportional to the concentration of holes [which is equivalent to the concentration of vacancies according to Eq. (6.52)], and we obtain a relationship: 1/8 σ ∝ [hž ] = K(T )1/4 PO2 (6.54) Experimentally, as illustrated in Figure 6.34, the electrical conductivity of Cu2 O is 1/7 found to be proportional to PO2 , which is in reasonable agreement with the prediction of Eq. (6.54). The variation of another oxide semiconductor, CdO, with temperature is also shown in Figure 6.34 for comparison. As a final example of oxide-based semiconductors, recall from Section 6.1.2.3 that some chalcogenides exhibit high ionic conductivities. Chalcogenides also play a role in the high conductivities found in some glasses (see Figure 6.35). In addition to the group VI elements, chalcogenide glasses sometimes contain elements of Group V, such as P, As, Sb, and Bi, as well as other ingredients like Ge or the halides. These conductivities are electronic, and not ionic, in nature. Their melting points are low, sometimes below 100◦ C, and they are quite sensitive to corrosion. The conductivity in these glasses can be explained in terms of a special type of glass structure defects around which pairs of positive and negative centers are present, and with which electrons couple. In addition to their semiconductivity, these glasses have a unique switching behavior. If a voltage is applied at the normal high-resistivity state, called the “off state,” a critical current–voltage pair of values is eventually achieved at which point the glass suddenly switches into a low-resistivity state, called the “on state.” Each state is stable, and the process is reversible. This is a type of storage, or memory, effect. 6.1.2.6 MOSFETs. A type of semiconductor device that utilizes oxide ceramics is a metal-oxide-semiconductor field-effect transistor, abbreviated as MOSFET. Just like the semiconductor junction devices of Section 6.1.1.6, the MOSFET is composed of nand p-type semiconductor regions within a single device, as illustrated in Figure 6.36.

Conductivity (Ω−1 • cm−1)

100

CdO, 600°C 10

Cu2O, 900°C

1 0.1

1 10 Oxygen pressure (mm Hg)

100

Figure 6.34 Electrical conductivity of two oxide semiconductors as a function of oxygen partial pressure. From W. D. Kingery H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

584

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

T 400°C 12

200

100

50

20

1

log r (r in Ω • cm)

9 2 3

6

4 5 3

6 0 1.5

2.0

2.5

3.0 × 10−3 K−1

3.5

1/T

Figure 6.35 Temperature dependence of resistivity for some semiconducting glasses: (1) vitreous silica; (2) soda–lime silicate glass; (3) AsSeTe chalcogenide glass; (4) silicate glass with 18 mol% (Fe3 O4 + MnO); and (6) borosilicate glass with 45 mol% (Fe3 O4 + MnO). Reprinted, by permission, from H. Scholze, Glass, p. 312. Copyright  1991 by Springer-Verlag.

Source

Gate

Drain

p -Type channel p -Type Si

SiO2 insulating layer

p -Type Si

n-Type Si substrate

Figure 6.36 Schematic cross-sectional view of a MOSFET transistor. Reprinted, by permission, from W. Callister, Materials Science and Engineering: An Introduction, 5th ed., p. 634. Copyright  2000 by John Wiley & Sons, Inc.

One such device consists of two small islands of p-type semiconductor with and ntype silicon substrate. The islands are joined by a narrow p-type channel. The oxide portion of the MOSFET is an insulating layer of silicon dioxide that is formed by surface oxidation of the silicon. Gate, drain, and source connectors are attached. The MOSFET differs from the junction transistor in that a single type of charge carrier, either an electron or a hole, is utilized, instead of both. The conductivity of the channel

ELECTRICAL PROPERTIES OF MATERIALS

585

is varied by the presence of an electric field imposed on the gate. For example, a positive field on the gate will drive holes out of the channel, thereby reducing the electrical conductivity. Thus, a small alteration in the field at the gate will produce a relatively large variation in current between the source and the drain. In this respect, the MOSFET is similar to the junction transistor, insofar as both create a signal amplification. The primary difference is that the gate current is exceedingly small in comparison to the base current of the junction transistor. As a result, MOSFETs are used where signal sources to be amplified cannot sustain an appreciable current. 6.1.3

Electrical Properties of Polymers

In this section, we consider the insulating properties of polymers as dielectrics, and we also describe their electrical- and ion-conducting capabilities. Given the highly insulative nature of most polymers, we begin with dielectric properties and then describe special types of polymers that conduct either electrons or ions. 6.1.3.1 Dielectric Polymers. Nonpolar polymers generally have dielectric constants in the 2–3 range while polar polymers can have values up to 7. Typical dielectric strengths of polymers are in the range of 20–50 kV/mm, which are approximately 2–10 times higher than ceramics and glasses and are hundreds of times higher than conducting metals and alloys. Data from mechanical and dielectric measurements can be related in a qualitative, if not quantitative, manner. Formally, the dielectric constant can be regarded as the equivalent of the mechanical compliance. This highlights the fact that the dynamic mechanical techniques of Section 5.3 measure the ability of a polymer to resist movement, whereas the dielectric technique measures the ability of the system to move, at least the dipolar portions. Interestingly, the dielectric loss appears to match the loss modulus more closely than the loss compliance when the data are compared for the same system. Both techniques respond in a similar fashion to a change in the frequency of the measurement. When the frequency is increased, the transitions and relaxations that are observed in a sample appear at higher temperatures. We expect maximum losses at transition temperatures for both electrical and mechanical deformation, as illustrated for poly(chlorotrifluoroethylene) in Figure 6.37. Just as mechanical fatigue after many alternating cycles can cause a polymer to fail at a lower stress, so dielectric insulation can fatigue under electrical stresses. 6.1.3.2 Electrically Conducting Polymers. High values of resistivity are common for organic polymers, with 1012 to 1018  · cm being typical. Electrical resistance can be lowered markedly by the addition of conductive fillers, such as carbon black or metal fillers, but the conductivity is due to the filler in this case. When we consider electronic conduction solely in homogeneous polymers, band theory is not totally suitable because the atoms are covalently bonded to one another, forming polymeric chains that experience weak intermolecular interactions. Thus, macroscopic conduction will require electron movement, not only along chains but also from one chain to another. Polymers that display electronic conductivity are usually insulators in the pure state but, when reacted with an oxidizing or reducing agent, can be converted into polymer salts with electrical conductivities comparable to metals. Some of these polymers are listed in Figure 6.38, along with the conductivities of metals and ceramics for

586

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Dielectric loss tangent × 102 and mechanical loss tangent × 10

5

4

Dielectric loss at 10 Hz

3

2

1

Mechanical loss at 3 to 12 Hz

0

−100°



100°

200°

Temperature, °C

Figure 6.37 Mechanical and dielectric loss tangents for poly (chlorotrifluoroethylene). Reprinted, by permission, from F. Rodriguez, Principles of Polymer Systems, 2nd ed., p. 271. Copyright  1982 by Hemisphere Publishing Corporation.

Organic polymers

Organic and molecular crystals

Inorganic materials

TMTSF2PF6 (superconducting) Cu 5

10

Conductivity (S cm−1)

100

Graphite Pyropolymers Polyacetylene Poly(p-phenylene) Poly(phenylene sulphide) Polypyrrole doped polymers

Hg TTF-TCNQ

Polythiophene

Cu-TCNQ

Polyphthalocyanine

Ag-TCNQ

10−20

Si

semiconductors

Boron H2O Iodine

Polyacetylene (undoped)

10−15

Bi Si ZnO Ge (doped)

10−5

10−10

metals

ZnO

Polydiacetylene Polythiophene Polypyrrole (undoped) Nylon Poly(p-phenylene) (undoped) Poly (phenylene sulphide) (undoped) PVC Polystyrene Polyimide (Kapton) PTFE (Teflon)

insulators Cu-phthalocyanine Anthracene

Diamond

SiO2

Figure 6.38 Conductivity ranges for doped and undoped polymers, inorganic materials and molecular crystals. One siemen (S) = 1 −1 . Reprinted, by permission, from J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed., p. 411. Copyright  1991 by J.M.G. Cowie.

ELECTRICAL PROPERTIES OF MATERIALS

587

comparison. The preparation and mechanism of conductivity in a few of these polymers will be elaborated upon. One of the first polymers to exhibit electrical conductivity was polyacetylene. Normally a very poor conductor in the pure state, it was found that a highly conductive polymer could be formed by reacting it with I2 . The result is a dramatic increase of over 1010 in conductivity. Polyacetylene is a polyconjugated polymer, the structure of which is shown in Table 6.10. Polyconjugation refers to the existence of multiple C–C bonds in the polymer, typically along the backbone. In a polyconjugated system, the π orbitals (cf. Section 1.0.4.2) are assumed to overlap, and they form a valance and a conduction band as predicted by band theory. If all the bond lengths were equal—that is, delocalization that led to each bond having equal double bond character—then the bands would overlap and the polymer would behave as a quasione-dimensional metal having good conductivity. Experimental evidence has shown that this is an unstable system that will undergo lattice distortion. The alternative is to have single and double bonds alternate along the backbone, which leads to an energy gap between the valence and conduction bands. The trans structure of polyacetylene has such an alternating single-bond–double-bond backbone. The trans structure of polyacetylene is also unique, because it has a twofold degenerate ground state in which sections A and B in Figure 6.39 are mirror images, and the single and double bonds can be interchanged without changing the energy. Thus, if the cis configuration begins to isomerize to the trans structure from different locations in a single chain, an Table 6.10 Structures and Conductivities of Doped Conjugated Polymers

Polymer

Typical Methods of Doping

Structure

Electrochemical, chemical (AsF5 , I2 , Li, K)

Polyacetylene n

Chemical (AsF5 , Li, K)

Polyphenylene

Typical Conductivity (S cm)−1 500–1.5 × 105 500

n

Poly(phenylene sulphide) Polypyrrole

Chemical (AsF5 )

S

1

n

N H

n

S

n

Polythiophene

Electrochemical

600

Electrochemical

100

C6H5

Poly(phenylquinoline)

N

Electrochemical, chemical (sodium naphthalide)

50

n

One siemen equals 1 −1 . Source: J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed. Copyright  1991 by J.M.G. Cowie.

588

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS •

A

Neutral soliton

B

Figure 6.39 Isomerization of cis (A) and trans (B) sequences in polyacetylene that meet to form a soliton. Reprinted, by permission, from J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed., p. 417. Copyright  1991 by J. M. G. Cowie.

A sequence may form and eventually meet a B sequence, as shown, but in doing so, a free radical, called a soliton, is produced. The soliton is a relatively stable electron with an unpaired spin and is located in a nonbonding state in the energy gap, midway between the conduction and valence bands. It is the presence of these neutral solitons which gives trans-polyacetylene the characteristics of an intrinsic semiconductor with conductivities of 10−7 to 10−8 ( · cm)−1 . The conductivity of polyacetylene can be magnified by doping. Exposure of a polyacetylene film to dry ammonia gas leads to a dramatic increase in conductivity of 103 ( · cm)−1 . Controlled addition of an acceptor, or p-doping, agent such as AsF5 , I2 , or HClO4 removes an electron and creates a positive soliton (or a neutral one if the electron removed is not the free electron). Similarly, a negative soliton can be formed by treating the polymer with a donor, or n-doping, agent that adds an electron to the mid-gap energy level. This can be done by an electrochemical method. At high doping levels, the soliton regions tend to overlap and create new mid-gap energy bands that may merge with the valence and conduction bands, allowing freedom for extensive electron flow. Thus, in polyacetylene the charged solitons are responsible for making the polymer a conductor. Poly(p-phenylene) has all the structural characteristics required of a potential polymer conductor. It is an insulator in the pure state, but can be doped using methods similar to those of polyacetylene to form both n- and p-type semiconductors. However, because poly(p-phenylene) has a higher ionization potential, it is more stable to oxidation and requires strong p-dopants. Examination of the structure shows that the soliton defect cannot be supported in poly(p-phenylene) because there is no degenerate ground state. Instead, it is assumed that conduction occurs because the mean free path of charge carriers extends over a large number of lattice sites and the residence time on any one site is small compared with the time it would take for a carrier to become localized. If, however, a carrier is trapped, it tends to polarize the local environment, which relaxes into a new equilibrium position. This deformed section of the lattice and the charge carrier then form a species called a polaron. Unlike the soliton, the polaron cannot move without first overcoming an energy barrier so movement is by a hopping motion. In poly(p-phenylene) the solitons are trapped by the changes in polymer structure because of the differences in energy and so a polaron is created which is an isolated charge carrier. A pair of these charges is called a bipolaron. In poly(p-phenylene) and most other polyconjugated conducting polymers the conduction occurs via the polaron or bipolaron (see Figure 6.40). There are other types of semiconducting polymers as well, some of the more important of which are listed in Table 6.10. The conduction mechanism in most of these polymers is the polaron model described above. Applications for these polymers are growing and include batteries, electromagnetic screening materials, and electronic devices.

ELECTRICAL PROPERTIES OF MATERIALS

589

Conduction bands − +

− −

Valence bands Neutral polymer Removal of has full valence one electron and empty forms polaron conduction bands

Removal of Removal of second electron more electrons forms bipolaron forms bipolaron bands

Figure 6.40 Proposed band structure for an oxidized p-type semiconducting polymer. Reprinted, by permission, from J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, 2nd ed., p. 419. Copyright  1991 by J. M. G. Cowie.

6.1.3.3 Ion-Conducting Polymers. Recall that electrical conductivity can be the result of electronic conductivity, in which electrons are the charge carriers, ionic conductivity, in which ions are the charges carriers, or a combination of both. In the previous section, we saw how some conjugated polymers can conduct electrons. In this section we concentrate on ion-conducting polymers. There are generally two types of polymers in this category: those that conduct primarily hydrogen ions and those that conduct primarily larger anions. In both cases, the term polymer electrolyte is used to describe the type of ion-transporting polymeric material, although it is more commonly used to describe the latter. This should not be confused with the term ionomer, which describes polymers that contain ionic groups primarily involved in creating ionic bonds between chains. This is not to say that ionomers are incapable of ionic conductivity—they may well be. It is rather a matter of terminology usage. Ionomers typically have applications other than as ion conductors. We begin by examining the structure and properties of polymer electrolytes. Initial investigations of ionically conductive polymers were principally focused on poly(ethylene oxide), PEO, when it was discovered that PEO can solvate a wide variety of salts, even at very high concentrations. The solvation occurs through the association of the metallic cations with the oxygen atoms in the polyether backbone (see Figure 6.41). The solvated ions can be quite mobile, and thus give rise to significant bulk ionic conductivities. Pure PEO is a semicrystalline polymer possessing both an amorphous and a crystalline phase at room temperature. Significant ionic transport occurs within the amorphous phase. This explains the dramatic decrease in ionic conductivity seen in many PEO-based systems for temperatures below the melting point of pure crystalline PEO(Tm = 66◦ C), as shown in Figure 6.42. The crystalline PEO regions are nonconductive and serve to hinder bulk ionic transport. The mobility of the ions in polymer electrolytes is linked to the local segmental mobility of the polymer chains. Significant ionic conductivity in these systems will occur only above the glass transition temperature of the amorphous phase, Tg . Therefore, one of the requirements for the polymeric solvent is a low glass-transition temperature; for example, Tg = −67◦ C for PEO. One problem associated with polymer electrolytes arises from the opposing effects of increasing the salt concentration. Higher salt concentrations generally imply a higher density of charge carriers, which increases conductivity. However, through their

590

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

carbon oxygen cation, e.g. Li+ anion, e.g. Br−

Figure 6.41 Schematic illustration of salt solvation by a PEO chain. Reprinted, by permission, from L. V. Interrante, L. A. Casper, and A. B. Ellis, eds. Materials Chemistry, p. 109. Copyright  1995 by the American Chemical Society.

T (C) 200

150

100

60 PbBr2 8:1 MgCl2 16:1 PEO PbI2 24:1 LiCF3SO3 9:1

Log s (Ω−1 • cm−1)

−4.0 −5.0 −6.0 −7.0 −8.0 −9.0 −10.0

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

3.4

1000/T (K−1)

Figure 6.42 Temperature dependence of conductivity for various PEO-based polymer electrolytes containing divalent cations. Reprinted, by permission, from L. V. Interrante, L. A. Casper, and A. B. Ellis, eds., Materials Chemistry, p. 110. Copyright  1995 by the American Chemical Society.

interactions with the polyether chains, inorganic salts increase the Tg as their concentrations increases. This increase in Tg tends to lower the mobility of the ions and thus decreases ionic conductivity. The balance of these two effects leads to a maximum in the conductivity at a specific salt concentration, which varies from system to system.

ELECTRICAL PROPERTIES OF MATERIALS

591

One method of reducing crystallinity in PEO-based systems is to synthesize polymers in which the lengths of the oxyethylene sequences are relatively short, such as through copolymerization. The most notable linear copolymer of this type is oxymethylene-linked poly(oxyethylene), commonly called amorphous PEO, or aPEO for short. Other notable polymer electrolytes are based upon polysiloxanes and polyphosphazenes. Polymer blends have also been used for these applications, such as PEO and poly(methyl methacrylate), PMMA. The general performance characteristics of the polymer electrolytes are to have ionic conductivities in the range of 10−3 ( · cm)−1 or (S/cm). Still higher total conductivities are achievable in a special class of hydrogen-ion (proton)-conducting polymers. These polymers are particularly useful for fuel-cell and chlor-alkali processing applications in which the efficient transfer of protons is critical. These polymers are of produced in membrane form and are therefore referred to as proton-exchange membranes (PEM). The most common polymeric proton-exchange membrane is Nafion , which is manufactured in various forms by Du Pont. The generalized chemical structure of Nafion is shown in Figure 6.43. It is a perfluorinated polymer, similar to Teflon , with small amounts of sulfonic (SO3 − ) or carboxylic functional groups. The microstructure of Nafion is actually quite complex. In general, the models, supported primarily by small-angle X-ray scattering (SAXS) results, predict that the ionic portions of Nafion tend to aggregate due to electrostatic interactions to form ionic clusters. It has been ˚ in size, resulting in (1) a determined that the clusters are on the order of 30–50 A phase-separated polymer with discrete hydrophobic regions formed from the polymer backbone and (2) hydrophilic regions formed from the ionic clusters. These distinct regions are functionally important. Nafion is believed to conduct protons due to extensive hydration that takes place in the ionic clusters. Waters of hydration are necessary, since according to one theory the protons are transported through the polymer in the form of hydronium ions, H3 O+ [5]. The proton conductivity of Nafion is orders of magnitudes higher than most other proton-conducting polymers (see Table 6.11). The conductivity, however, is CF2

CF2

x

CF OCF2

CF2

y

CF

z

O(CF2)2

SO3−H+

CF3

Figure 6.43

The generalized chemical structure of Nafion .

Table 6.11 Proton Conductivities of Selected Polymers Polymer

Conductivity ( · cm)−1

Nafion

2.2 × 10−1

Polybenzimidazole (PBI) Poly(vinylidine)fluoride

2 × 10−4 –8 × 10−4 10−2

592

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

highly dependent upon (a) the relative humidity in the membrane and (b) the ability of the membrane to stay hydrated during operation. One particular drawback to all tetrafluoroethylene-based polymers is the inherent difficulty in processing due to limited solubility in most solvents. As a result, only certain types of geometries, such as thick films, are commercially available. 6.1.4

Electrical Properties of Composites

Because electronic and ionic conduction are so structure-sensitive, the simple ruleof-mixtures approach to estimating the conductivity and resistivity of composites is not normally of use. As a result, the conductivity of specific composites for specific applications must be experimentally determined. In the next two sections, we examine two examples of how composites can be used in electrical applications, and we describe the influence of each component on the electrical properties. The first example involves the electrical insulating properties of polymers, and the second one involves enhancing the electrically conducting properties of polymers. 6.1.4.1 Dielectric Properties of Glass-Fiber-Reinforced Polymers. As described in several of the previous sections on composites, the most common type of glass fiber used for reinforcement purposes is E-glass, which was originally developed for electrical insulation purposes, hence its name (“E” stands for “electrical grade”). It is not surprising, then, that when combined with organic resins such as epoxy, a class of materials results which can be readily molded into complex shapes that possess excellent insulting properties and high dielectric strength. Some electrical properties of reinforcing fibers, composite resins, and the resulting composites are given in Tables 6.12, 6.13, and 6.14, respectively. These values should be taken as approximate only, especially for the composites, since fiber orientation, content, and field strength have an enormous impact on the dielectric properties of these materials. Some of the most widespread electrical applications for glass-fiberreinforced epoxy systems are in printed circuit boards and electrical housing such as junction boxes. One interesting application of the dielectric properties of the resins in these composite systems is that it provides an opportunity to monitor the cure characteristics of

Table 6.12 Some Electrical Properties of Selected Reinforcing Glass Fibers Property Dielectric constant, 106 Hz Dielectric constant, 1010 Hz Loss tangent, 106 Hz Loss tangent, 1010 Hz

E-Glass

S-Glass

D-Glass

5.80

4.53

3.56

6.13 0.001 0.0039

5.21 0.002 0.0068

4.00 0.0005 0.0026

Source: Handbook of Fiberglass and Advanced Plastics Composites, G. Lubin, ed. Copyright  1969 by Reinhold Book Corporation.

ELECTRICAL PROPERTIES OF MATERIALS

593

Table 6.13 Some Electrical Properties of Selected Resins Property

Phenolic

Epoxy

Polyester

Silicone

1012 –1013

1016 –1017

1014

1011 –1013

Dielectric strength, V/µm

14–16

16–20

15–20

7.3

Dielectric constant, 60 Hz

6.5–7.5

3.8

3.0–4.4

4.0–5.0

Resistivity,  · cm

Source: Handbook of Fiberglass and Advanced Plastics Composites, G. Lubin, ed. Copyright  1969 by Reinhold Book Corporation.

Table 6.14 Some Electrical Properties of Selected Glass-Fiber-Reinforced Resins

Property

GFR Nylon

GFR GFR Polypropylene Phenolic

Resistivity,  · cm 2.6–5.5 × 1015

1.7 × 1016

GFR (Laminate)

GFR GFR Silicone Polyester (Laminate)

1010 –1011 6.6 × 107 –109 1012 –1013 2–5 × 1014

Dielectric strength, V/µm

16–18

12–19

15–17

26–30

8–16

28.5

Dielectric constant, 60 Hz

4.6–5.6

2.3–2.5

7.1–7.2

4.4

4.6–5.2

3.9–4.2

Source: Shackelford, J. and W. Alexander, The CRC Materials Science and Engineering Handbook. Copyright  1992 by CRC Press.

Cooperative Learning Exercise 6.4

Assume that the conductivity of a undirectional, continuous fiber-reinforced composite is a summation effect just like elastic modulus and tensile strength; that is, an equation analogous to Eq. (5.88) can be used to describe the conductivity in the axial direction, and one analogous to (5.92) can be used for the transverse direction, where the modulus is replaced with the corresponding conductivity of the fiber and matrix phase. Perform the following calculations for an aluminum matrix composite reinforced with 40 vol% continuous, unidirectional Al2 O3 fibers. Use average conductivity values from Appendix 8. Person 1: Calculate the conductivity of the composite in the axial direction. Person 2: Calculate the conductivity of the composite in the transverse direction. Compare your answers. How many orders of magnitude difference is there in the two conductivities? Answers: σaxial = (0.4)(10−11 ) + (0.6)(3.45 × 107 ) = 20.7 × 106 (-m)−1 ; σtrans = (3.45 × 107 )(10−11 )/[(0.4)(3.45 × 107 ) + (0.6)(10−11 )] = 2.5 × 10−11 (-m)−1 . 17 orders.

594

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

the resin in the composite. In-process cure monitoring ensures that each molded part has properly cured before the mold is opened. The dielectric loss [see Eq. (6.47)] has been shown to increase rapidly in the beginning of the cure cycle (see Figure 6.44), attain a peak, then reduce to a constant value as the curing reactions are completed. At the beginning of the cure cycle, the resin viscosity in the uncured prepreg is relatively high so that the dipole and ion mobilities are restricted. This results in low loss factors just after the uncured prepreg is placed in the mold. As the temperature of the prepreg increases in the mold, the resin viscosity is reduced and the loss factor increases owing to greater dipole and ion mobilities (see Figure 6.25). As soon as the gel point is reached, the resin viscosity increases rapidly and the loss factor decreases. At a full degree of cure, the loss factor levels off to a constant value. The technique for monitoring the dielectric loss factor is relatively simple. Two metal electrodes are placed opposite each other at critical locations on opposite sides of the mold. When the sheet molding compound (SMC), is placed between the electrodes, a capacitor is formed. The dielectric power loss is monitored continually throughout the molding cycle, as outlined in Section 6.1.2.2. 6.1.4.2 Metal Oxide-Polymer Thermistors. The variation of electrical properties with temperature heretofore described can be used to tremendous advantage. These so-called thermoelectric effects are commonly used in the operation of electronic temperature measuring devices such as thermocouples, thermistors, and resistancetemperature detectors (RTDs). A thermocouple consists of two dissimilar metals joined at one end. As one end of the thermocouple is heated or cooled, electrons diffuse toward

400

4 TEMPERATURE 3

2

1Hz 200

1 100 Hz

TEMPERATURE (° F)

LOG LOSS FACTOR

300

100 0

−1

10000 Hz

0

60

120

180 TIME (MINUTES)

240

0 300

Figure 6.44 Dielectric loss factor as a function of cure time and frequency of the oscillating electric field in a fiber-reinforced polymer. Reprinted, by permission, from P. K. Mallick, Fiber-Reinforced Composites, p. 365. Copyright  1988 by Marcel Dekker, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

595

LOG RESISTIVITY (Ω • Cm)

the cold end, and a Seebeck voltage is produced in each wire. Owing to the intentional dissimilarity between the Seebeck voltage in the two wires that compose the thermocouple, a voltage difference is established between the two wires, which can be measured and calibrated to temperature. A thermistor, on the other hand, directly correlates the change in resistivity of a substance to temperature. Although thermocouples are more widely utilized due to their inherent stability and larger operating temperature range, the thermistor concepts allows us to illustrate a thermoelectric effect that utilizes composite materials. Thermistors can exhibit two primary resistance effects (see Figure 6.45): the negative temperature coefficient (NTC) effect, and the positive temperature coefficient (PTC) effect. In composites, these transitions arise from an individual component, such as a semiconductor–metal transition exhibited by V2 O3 metallic filler, or result from a product property of the composite. Traditionally, PTC thermistors have been prepared from doped-BaTiO3 , such that the grains are semiconducting and the grain boundaries are insulating. The PTC effect occurs during heating through the Curie point, at which point the crystal structure changes from tetragonal to cubic, accompanied by a ferroelectric transition (see Section 6.1.2.2). In the ferroelectric state, the spontaneous polarization compensates for the insulating grain boundaries, resulting in a low resistivity. Although BaTiO3 thermistors exhibit relatively large (six orders of magnitude) changes in resistivity with temperature, they have high room-temperature resistivities and are expensive to produce. Composite-based PTC thermistors are potentially more economical. These devices are based on a combination of a conductor in a semicrystalline polymer—for example, carbon black in polyethylene. Other fillers include copper, iron, and silver. Important filler parameters in addition to conductivity include particle size, distribution, morphology, surface energy, oxidation state, and thermal expansion coefficient. Important polymer matrix characteristics in addition to conductivity include the glass transition temperature, Tg , and thermal expansion coefficient. Interfacial effects are extremely important in these materials and can influence the ultimate electrical properties of the composite.

PTC NTC

TEMPERATURE (°C)

Figure 6.45 Schematic illustration of negative temperature coefficient (NTC) and positive temperature coefficient (PTC) effects. Reprinted, by permission, from D. M. Moffatt, J. Runt, W. Huebner, S. Yoshikawa, and R. Nenham, in Composite Applications, T. L. Vigo and B. J. Kinzig, eds., p. 52. Copyright  1992 by VCH Publishing, Inc.

596

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

There is a relatively sharp increase in resistivity near the melting temperature of the polymer, Tm , at which point a discontinuous change in specific volume occurs. The PTC phenomenon in these composites is believed to be the result of a separation of the conducting particles due to the thermal expansion of the polymer. Because the transition occurs at the melting point, one may control the PTC transition by utilizing a different polymer matrix. Composites prepared with crystalline polymers can also exhibit dramatic NTC effects just above the melting point, due to the relaxation of the polymer structure. This effect may be eliminated by crosslinking the polymer or adding a third filler to stabilize the matrix. Although a majority of these composite thermistors are based upon carbon black as the conductive filler, it is difficult to control in terms of particle size, distribution, and morphology. One alternative is to use transition metal oxides such as TiO, VO2 , and V2 O3 as the filler. An advantage of using a ceramic material is that it is possible to easily control critical parameters such as particle size and shape. Typical polymer matrix materials include poly(methyl methacrylate) PMMA, epoxy, silicone elastomer, polyurethane, polycarbonate, and polystyrene. The resistivity–temperature behavior for a V2 O3 –PMMA composite is shown in Figure 6.46. When the NTC transition in the V2 O3 is combined with the PTC effect in the composite, a resistivity–temperature curve with a square-well appearance results. As the amount of V2 O3 is increased, the PTC effect begins to appear near 50 vol% V2 O3 . The optimum composition is that which yields the lowest room-temperature resistivity and the largest PTC effect. As the amount of conducting filler is increased, the temperature at which the PTC effect occurs moves to higher temperatures. The PTC exhibited in the composite is related to a phenomenon called percolation, which deals with the number and properties of neighboring entities and the existence of a sharp transition at which long-range connectivity in the system appears or disappears. At low

12

LOG RESISTIVITY (Ω • Cm)

11

10,20%

10

30,40 %

9 8 7

50%

6

55%

5 4 3

60% 65%

2 1 −200

0

200

TEMPERATURE (°C)

Figure 6.46 Temperature dependence of resistivity for a V2 O3 –PMMA composite of various volume fractions V2 O3 at 1 kHz. Reprinted, by permission, from D. M. Moffatt, J. Runt, W. Huebner, S. Yoshikawa, and R. Nenham, in Composite Applications, T. L. Vigo and B. J. Kinzig, eds., p. 56. Copyright  1992 by VCH Publishers, Inc.

ELECTRICAL PROPERTIES OF MATERIALS

597

filler concentrations, a high resistivity is due to the insulating phase. As the volume fraction is increased, conducting paths start to form, resulting in a decrease in resistivity. This region is known as the percolation threshold. At higher filler concentrations, a saturation region is reached wherein the resistivity is relatively constant due to extensive particle contacts and eventually approaches the value of the pure conductor. 6.1.5

Electrical Properties of Biologics

The principles introduced in the previous sections have important applications in the description of many biological processes. Both ionic and electron charge transfer play important roles in bioprocesses such as nerve and muscle cell function. While an understanding of these processes would be a worthwhile academic exercise, there is little new to be learned about the materials involved—that is, structure–property relationships for the biological materials. What would be useful to us is a description of the electrical properties of the biological materials—for example, the electrical conductivity of various tissues, or the ion-conducting capacity of membranes. Unfortunately, while some of this information may indeed exist, it is difficult to find all the necessary information, and it is even more difficult to make generalizations. In other words, the human body, and other biological units like it, are complex chemoelectric devices that are not yet fully understood from a materials standpoint. There are any number of ongoing scientific studies aimed at better understanding these biological processes and utilizing them for the development of biomaterials, such as the stimulation of nerve cell growth using electrically conducting polymers [6], the electrical properties of amphiphilic lipid membranes [7], and the electrical properties of hydroxyapatite thin films [8]. These topics, while fascinating, are much too specific for this introductory text. Instead, let us concentrate on two, more general, classes of artificial biomedical devices that may give us some insight into the electrical processes in biologics: biosensors and bioelectrodes. 6.1.5.1 Biosensors. The term bioelectrode is broadly used to denote a class of devices that transmit information into or out of the body as an electrical signal. Those bioelectrodes that transmit information out of the body are generally called biosensors. Any type of sensor falls into one of two general categories: physical or chemical. Physical parameters of biological interest include pressure, volume, flow, electrical potential, and temperature. Chemical sensing involves the determination of the concentration of chemical species in a volume of gas, liquid, or tissue. The species can vary in size from a hydrogen ion to a live pathogen. When the species is complex, such as a parasite or a tumor, an interaction with another biological entity is required to recognize it. In this section, we give two examples of chemical biosensors: ion sensors and glucose sensors. Many simple ions such as K+ , Na+ , Cl− , and Ca2+ are normally kept within a narrow range of concentrations in the body, and they must be monitored during critical care. Potentiometric sensors for ion, also called ion-selective electrodes or ISEs, utilize a membrane that is primarily semipermeable to one ionic species. The ionic species is used to generate a voltage that generally obeys the Nernst equation [cf. Section 3.1.3.2 and Eq. (3.24)] RT ln[ai + kij (aj )z/y ] (6.55) E=C+ zF

598

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

where E is the potential in response to an ion, i, of activity ai and charge z; C is a constant; kij is the selectivity coefficient; and j is any interfering ion of charge y and activity aj . Glasses exist that function as selective electrodes for many different monovalent and some divalent cations. Alternatively, a hydrophobic membrane can be made semipermeable if a hydrophobic molecule called an ionophore that selectively binds an ion is dissolved in it. The selectivity of the membrane is determined by the structure of the ionophore. Some ionophores are natural products, such as gramicidin, which is highly specific for K+ , whereas others such as crown ethers and cryptands are synthetic. Ions such as S2− , I− , Br− , and NO− 3 can be detected using quaternary ammonium cationic surfactants as a lipid-soluble counterion. ISEs are generally sensitive in the 10−1 to 10−5 M range, but are not perfectly selective. The most typical membrane material used in ISEs is polyvinyl chloride plasticized with dialkylsebacate or other hydrophobic chemicals. Biosensors are also available for glucose, lactate, alcohol, sucrose, galactose, uric acid, alpha amylase, choline, and L-lysine. All are amperometric sensors based on O2 consumption or H2 O2 production in conjunction with the turnover of an enzyme in the presence of substrate. In the case of glucose oxidase reaction, the normal biological reaction is: Gluconic acid + H2 O2 (6.56) Glucose + O2 + H2 O Under many circumstances, the concentration of oxygen is rate-limiting, so the sensor often measures not glucose, but the rate at which oxygen diffuses to the enzyme to reoxidize its cofactor. There are two electrochemical ways to couple the reaction to electrodes: (a) monitor depletion of oxygen by reducing what is left at an electrode or (b) monitoring buildup of H2 O2 by oxidizing it to O2 and protons: H2 O2

O2 + 2H+ + 2e−

(6.57)

An alternative method is to use electrochemical mediators that are at a higher concentration that O2 and can therefore be shuttled back and forth between the protein and the electrode faster than the enzyme is reduced, so that the arrival of the glucose is always rate-limiting. A typical chemical that works in this way is ferrocene, which is an iron cation between two cyclopentadienyl anions, as shown in Figure 6.47. It exists in neutral and +1 oxidation state that are readily interconvertible at metal or carbon electrodes. Most of the sensors are macroscopic and are employed in the controlled environment of a clinical chemistry analyzer. However, complete glucose sensors are available that

Fe

Figure 6.47 The structure of ferrocene.

ELECTRICAL PROPERTIES OF MATERIALS

599

contain disposable glucose oxidase-based electrodes, power supply, electronics, and readout in a housing the size of a ballpoint pen. The ultimate goal is an implantable glucose sensor, toward which much progress is being made. 6.1.5.2 Electrical Stimulation Devices. Bioelectrodes that transmit electrical signals into the body are generally known as electrical stimulation devices, examples of which include cardiac pacemakers, transcutaneous electronic nerve stimulators (TENs) for pain suppression, and neural prostheses such as auditory stimulation systems for the deaf and phrenic nerve stimulators for artificial respiratory control. In these, and other similar devices, electrodes transmit current to appropriate areas of the body for direct control of, or indirect influence over, target cells. Electrically excitable cells, such as those of the nervous system or heart, possess a potential difference across their membranes of approximately 60–90 mV. The inside of the cell is negative with respect to the outside. Such cells are capable of transmitting electrical signals, called action potentials, along their lengths. During an action potential, the cellular membrane (cf. Figure 1.89) changes polarity for a duration of about 1 ms, so that the inside of the cell becomes positive with respect to the outside. Excitable cells can be artificially stimulated by electrodes that introduce a transient electric field of proper magnitude and distribution. In general, the field generated near a cathodically driven electrode can be used to depolarize most efficiently an adjacent excitable cell’s axonal process above a threshold value at which an artificially generated action potential results. Cathodes will tend to draw current outward through nearby cell membranes. In their passive state, such membranes can be electrically modeled as parallel arrangements of resistors and capacitors. Outward current therefore elicits depolarizing resistive potential drops and capacitive charging. Depending upon the specific application and electrode properties, a single action potential in a nerve cell might be elicited using currents on the order of a few microamps to a few milliamps, for durations of micro- to milliseconds. The materials used in bioelectrodes for electrical stimulation vary, depending upon the specific application. For example, intramuscular stimulation requires an electrode with excellent long-term strength and flexibility, but not very high stimulation charge densities. Stimulation of the visual cortex through placement of electrodes on the cortical surface, however, requires little mechanical flexibility in an electrode material with relatively moderate charge needs. Intracortical stimulation electrodes must be fabricated on a very small scale and must be capable of injecting very high charge densities. In all instances, the device must be biocompatible. In most neural control devices, the surface area of the exposed material is relatively small (< 1 cm2 ) so that problems associated with passive biocompatibility are minor. The noble metals such as platinum, iridium, rhodium, gold, and palladium are generally preferred for the conductor material due to their low chemical reactivity and high resistance to corrosion. The insulating materials around the conductive electrodes, however, have a much higher volume and surface area, and they represent the largest problems associated with biocompatibility. Thin layers of medical-grade silicone rubber, poly(tetrafluoroethylene), polyimide, or epoxy are typically used for this purpose. Other devices are fabricated using lithography and thin-film technology that are typically insulated with ceramics such as SiO2 and Si3 N4 . These insulting materials must be carefully selected to provide pinhole-free coatings, good adhesion to the conductors, and biocompatibility with the tissue. Once inserted, charge transfer between the metal electrode and biological tissue requires a change in charge carriers from electrons in the metal to ions in the tissue

600

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

fluid. This change in charge carriers occurs by one of two mechanisms. The first is a capacitive mechanism involving only the alignment of charged species at the electrode–tissue interface, in which charging and discharging occurs at the electrode double layer. This mechanism is preferred because no chemical change occurs in either the tissue or the electrode. The amount of charge that can be transferred solely by capacitive charging is only about 20 µC/cm2 of “real” electrode surface area, which is the actual area of the electrode–tissue interface. The second mechanism involves the exchange of electrons across the electrode–tissue interface and therefore requires that some chemical species be oxidized or reduced. Metal electrodes almost always inject charge by this so-called faradaic charge transfer process, because the amount of charge required greatly exceeds that available from the capacitive mechanism alone. The types of reactions that occur at the electrode–tissue interface include monolayer oxide formation and reduction, the electrolysis of water, or the oxidation of chloride ions. Capacitor electrodes are considered the ideal type of stimulation electrode because the introduction of a dielectric at the electrode–tissue interface allows charge flow completely by charging and discharging of the dielectric film. The oxide film that is produced on tantalum by anodic polarization withstands substantial voltage without significant current leakage and permits charge flow without the risk of faradaic charge transfer reactions. Anodized titanium has higher dielectric strength than anodized tantalum, but its current leakage is too high for electrode applications. Capacitor electrodes based on tantalum–tantalum pentoxide have a charge storage capacity of about 100–150 µC/cm2 . Their use is limited to neural prosthesis applications having electrodes about 0.05 cm2 or larger in surface area, such as peripheral nerve stimulators. It is not possible to obtain adequate charge storage capacity in electrodes smaller then 10−3 cm2 .

6.2

MAGNETIC PROPERTIES OF MATERIALS

Just as materials have a response when placed in an electric field, they can have a response when placed in a magnetic field. We will see in this section that many of the concepts of permanent dipoles and dipole alignment in response to an applied field that were described in the context of electrical fields apply to magnetic fields as well. There are a few differences, however, and we will also see that there are fewer materials with specialized magnetic properties than there were with specialized electrical and electronic properties. The magnetic properties of materials are nonetheless important, and they are applied in a number of technologically important areas. 6.2.1

Magnetic Properties of Metals and Alloys

6.2.1.1 The Molecular Origins of Magnetism. Recall from Section 1.0.3.2 that charge separation in a molecule leads to an electric dipole, and also recall that this dipole can be either permanent (intrinsic) or induced. And just as the magnitude of the electric dipole moment, pe , was proportional to the charge and separation distance, so is the magnitude of the magnetic dipole moment, pm , proportional to the pole strength, m, and the separation distance, a:

pm = m · a

(6.58)

MAGNETIC PROPERTIES OF MATERIALS

601

Intrinsic magnetic dipole moments arise from two sources: (a) the orbital motion of electrons about the nucleus and (b) the net spin of unpaired electrons. In the first case, the intrinsic magnetic dipole is not observed until acted on by an external field. In the presence of an external magnetic field, the component of the magnetic moment that is parallel to the applied field is given by ml (eh/4πme ), where ml is the magnetic quantum number, e and me are the magnitude of the electronic charge and mass, respectively, and h is Planck’s constant. Any filled shell or subshell contributes nothing to the orbital magnetic moment of an atom, but occupied outer electronic orbitals contribute to a net atomic magnetic moment (parallel to the applied field) if the summation of magnetic quantum numbers is not equal to zero. The quantity eh/4πme is a fundamental quantity in magnetism called the Bohr magneton and is assigned the symbol µB . The Bohr magneton has a value of 9.27 × 10−24 A · m2 . The quantity h/2π is the angular momentum of the electron. The second type of intrinsic magnetic dipole moment is associated with the spin of an electron. When the spin quantum number, ms , is + 1/2 (spin up), the magnetic moment is +µB (parallel to the magnetic field), and when ms = − 1/2 (spin down), the magnetic moment is −µB (antiparallel to the magnetic field). For free atoms like Na and Mg, no permanent magnetic moment exists because there are no unpaired electrons and the individual orbital momentums sum to zero. Free atoms like Na and molecules like O2 , each of which have unpaired electrons (see Figure 1.7), possess a permanent magnetic dipole moment. When both spin and orbital components of the magnetic dipole moment exist, they are additive, though not in a simple arithmetic manner. Recall from Section 1.0.4.2 that the pairing of electrons in molecular orbitals of molecules results in two types of magnetism. Diamagnetism is the result of all paired electrons in the molecular orbitals, whereas paramagnetism results from unpaired electrons in the molecular orbitals. Let us now examine how these two types of magnetism behave in the presence of an imposed magnetic field. This can be done by comparing the magnetic induction, B, or internal magnetic field strength, within a substance in the presence of an external magnetic field, H , to the induction produced by the field alone in a vacuum, B0 . First, consider the case of a field in a vacuum. The field can be created by passing current, I , through a coil of N turns (called a solenoid ) of total length, l, as shown in Figure 6.48. The field is simply given by: H =

N ·I l

(6.59)

Recall from Eq. (6.2) that the induction and field are related as follows: B0 = µ0 H

(6.60)

where the subscript “0” has been added to indicate a vacuum, and boldface has been dropped to indicate that the inductance and field—normally vectors—are treated as scalar quantities only. The constant µ0 is the magnetic permeability in vacuum and has a value of 4π × 10−7 Wb/A · m. If a substance is now placed inside the solenoid, as shown in Figure 6.49, the magnetic field remains the same according to Eq. (6.59) as long as the current remains

602

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

I

B0 = m0H

H

N turns

I

Figure 6.48 Schematic diagram of a solenoid in a vacuum. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

I

B = mH

H

I

Figure 6.49 Schematic diagram of a substance within a solenoid. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

the same, but due to interactions between the field and the substance, the inductance is different (lower or higher), and Eq. (6.60) becomes B = µ0 H + µ0 M

(6.61)

where the second term contains the magnetization, M, of the substance. Magnetization represents the net magnetic dipole moment per unit volume that is aligned parallel

MAGNETIC PROPERTIES OF MATERIALS

603

to the external field. For many substances, a direct proportionality exists between the magnetization and the magnetic field: M = χH

(6.62)

with the dimensionless proportionality constant, χ, called the magnetic susceptibility. Do not confuse the magnetic susceptibility with the electronegativity of an element, which is represented by the same variable. The benefit of introducing magnetic susceptibility is that it is a material property, independent of quantity. Even if a direct proportionality does not exist between the magnetization and the magnetic field, Eq. (6.62) still defines the instantaneous magnetic susceptibility and can be substituted into Eq. (6.61) to yield B = µ0 (1 + χ)H (6.63) The quantity (1 + χ) is called the relative magnetic permeability (also dimensionless), which is also a material property.

Cooperative Learning Exercise 6.5

A coil of wire 0.20 m long and having 200 turns carries a current of 10 A. Person 1: Calculate the magnetic field strength in vacuum. Person 2: Look up the magnetic susceptibility for titanium. A bar of titanium is now placed in the coil. Combine your information to make the following calculations. Person 1: Calculate the flux density (magnetic induction) with the titanium bar in the coil. Person 2: Calculate the flux density (magnetic induction) in vacuum (without the titanium bar). Compare your answers. Does the titanium bar affect the magnetic induction significantly? Why or why not? Person 1: What is the relative magnetic permeability of titanium? Person 2: What is the magnetization of titanium inside the coil? Answers: Eq. (6.59) H = (200)(10)/(0.2) = 10,000 A/m; from Table 6.15 χ = 0.182 × 10−3 . Eq. (6.63) B = 4π × 10−7 (1 + 0.182 × 10−3 )(10, 000) = 0.0126 Wb/m2 ; B0 = 4π × 10−7 (10, 000) = 0.0125 Wb/m2 ; they are essentially the same due to the small susceptibility of titanium. (1 + χ) = 1.000182; Eq. (6.62) M = (0.000182)(10, 000) = 1.82 A/m.

With this background on magnetic field production and susceptibility in hand, we can return to our description of diamagnetic and paramagnetic behavior in an applied magnetic field. Substances consisting of atoms or molecules that do not possess permanent magnetic moments exhibit diamagnetism in a magnetic field. That is, they display a negative, but generally quite small, magnetic susceptibility because of the weak, induced opposing magnetic dipole moments. An induced diamagnetic component of magnetization is inherent in all substances, but is often masked by larger intrinsic moments that reinforce an applied field. Inert gases and solids like Bi, Cu, MgO, and

604

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

diamond are examples of diamagnetic materials and have typical magnetic susceptibilities on the order of −10−5 . The diamagnetic component of susceptibility is independent of temperature and magnetic field. For the most part, diamagnetic materials are of little engineering significance, with the notable exception of superconductors. Type I superconductors (cf. Section 6.1.2.4) can be perfectly diamagnetic (χ = −1 or B = 0) up to the critical magnetic field because induced shielding currents persist indefinitely. Type II superconductors exhibit perfect diamagnetism or partial paramagnetism depending on the applied field (cf. Figure 6.31). The strong diamagnetism of superconductors makes it possible to use such materials for magnetic field shielding purposes. Substances having magnetic susceptibilities in the range from about 10−6 to 10−2 are called paramagnetic. As described above, paramagnetism results from net intrinsic magnetic moments associated with electronic orbitals and electron spins. For paramagnetic materials, the interactions between atomic moments and an applied field can be described in terms of thermodynamics. When the individual atomic moments are randomly arrayed, the entropy of the material is high. Countering this tendency is the magnetic interaction energy, which is decreased when the moments align with the field. The final degree of moment alignment, therefore, depends on the competing factors of energy and entropy. Thus, the susceptibility, which reflects the degree of alignment, tends to decrease as temperature increases and entropy becomes more dominant. The results of thermodynamic calculations that minimize the free energy as a function of temperature, and they show that the susceptibility decreases with temperature and at intermediate temperatures is given approximately by χ=

2 µ0 Npm 3kB T

(6.64)

where pm is the magnetic dipole moment per atom as given in Eq. (6.58), N is the number of atoms per unit volume, kB is Boltzmann’s constant, T is absolute temperature, and µ0 is the magnetic permeability in vacuum constant. The dependence of magnetic susceptibility on reciprocal absolute temperature is called the Curie law, and it holds for most paramagnetic nonmetallic materials above room temperature. As shown in Figure 6.50, the Curie Law does not hold at low temperatures, and the susceptibility approaches a constant value due to saturation. In metals, conduction electrons provide both paramagnetic and diamagnetic moments, with a net paramagnetic moment, in addition to the induced diamagnetic moment resulting from core electrons. The paramagnetic component arising from the conduction electrons usually predominates, and most metals have positive susceptibilities. The origin of this paramagnetism can be explained in terms of the conduction electron energy band viewed as two subbands, one containing electrons with spin up and the other containing valence electrons with spin down. With no applied magnetic field, the subbands have equal populations of electrons and there is no net spin. Application of a magnetic field lowers the potential energy for the subband having spins parallel to the field and raises the potential energy of the other subband. Since the Fermi energy must be at the same level in both subbands when equilibrium exists (i.e., the electronic free energy is constant), a greater number of conduction electrons reinforce the field than oppose it, as illustrated in Figure 6.51. Those electrons with moments aligned parallel to the field (ms = + 1/2) have a magnetic potential energy 2µ0 µB H lower than those electrons with moments antiparallel (ms = − 1/2).

605

χ

MAGNETIC PROPERTIES OF MATERIALS

χ∼ 1 T (high temperatures)

T

Figure 6.50 Temperature dependence of magnetic susceptibility for a paramagnetic material. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

E EF Moments parallel to field

Moments antiparallel to field

2 m0mBH

0

N (E ) (for mS = +

1 ) 2

N (E ) (for mS = − 1 ) 2

Figure 6.51 Distribution of electrons within the conduction band of a metal in the presence of an applied magnetic field. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

606

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Equalization of the Fermi energy within these subbands leads to a greater population of electrons with moments parallel to the field and a resulting net magnetic moment. It can be shown that the conduction electron net spin susceptibility is proportional to the temperature coefficient of the electronic heat capacity [cf. Eq. (4.42)] and, for free electrons in a single band, having the Fermi energy much lower than any band gap, is given by µ0 Nv µ2B (6.65) χ= EF where EF is the energy of the Fermi level, Nv is the total number of valence electrons per unit volume [see Eq. (6.23)], µ0 is the magnetic permeability in vacuum constant, and µB is the Bohr magneton. Figure 6.51 shows us that only those electrons having energies near the Fermi level are responsible for spin paramagnetism, and this has been accounted for in Eq. (6.65). In general, the spin susceptibility is proportional to the density of states at the Fermi energy, so that transition metals exhibit larger paramagnetic susceptibilities than do nontransition metals (see Table 6.15), just as they exhibit larger electronic heat capacities. Some metals are diamagnetic because the conduction electron spin susceptibility is smaller than the induced diamagnetic susceptibility component. On the other hand, various rare earth metals display very strong paramagnetism because of unpaired f electrons that remain associated with individual atoms rather than entering into energy bands. 6.2.1.2 Types of Magnetism. In addition to diamagnetic and paramagnetic behavior, there are other types of magnetism which are important. These include ferromagnetism, antiferromagnetism, and ferrimagnetism. The various types of magnetism will be summarized and compared here, and then described in more detail under the appropriate material-related section. In most cases, the different types of magnetism can be differentiated between by their variation of magnetic susceptibility with temperature. The elements Fe, Co, Ni, Gd, Tb, Dy, Ho, and Tm, as well as a number of alloys and compounds, exhibit very large positive susceptibilities below a critical temperature, θc , known as the ferromagnetic Curie temperature. (We use the Greek lowercase theta and subscript “c” to designate the Curie temperature to differentiate it from the critical transition temperatures of superconductors, which are designated as Tc .) Such substances display spontaneous magnetization, the origin of which arises from an internal interaction that causes a permanent offset in the electronic subbands. Only certain metals and other solids having partially filled d subshells or partially filled f subshells are capable of ferromagnetism (see Table 6.16). In addition to this necessary, but not sufficient, condition for spontaneous permanent magnetization, the interatomic spacing must be such that an internal interaction, termed the exchange interaction, occurs. As shown in Table 6.16, the elements Fe, Co, and Ni are magnetized spontaneously, whereas Cu does not. This is because in Fe, Co, and Ni, the average number of electrons in the d band having spin up is not the same as that having spin down, and the difference between the two gives a net magnetic moment. With increasing temperature, the lowering of internal energy because of the exchange interaction is opposed by a greater entropy that results from the disorder of randomized spins. Ferromagnetic materials become paramagnetic above θc ; susceptibilities can be on the order of 250 to 100,000 or more and continue to increase with decreasing temperature

607

Mg 0.012

Ca 0.019

Sr 0.034

Ba 0.007

Na 0.008

K 0.006

Rb 0.004

Cs 0.005

Rare Earths

Y 0.114

Sc 0.264

Al 0.021

IIIA

Hf 0.070

Zr 0.109

Ti 0.182

IVA

Ta 0.178

Nb 0.226

V 0.375

VA

W 0.078

Mo 0.119

Cr 0.313

VIA

Re 0.094

Tc 0.395

Mn 0.871

VIIA

Os 0.015

Ru 0.066

Fe Ferromagnetic

Ir 0.038

Rh 0.169

Co Ferromagnetic

VIII

Pt 0.279

Pd 0.802

Ni Ferromagnetic

Au Diamagnetic

Ag Diamagnetic

Cu Diamagnetic

IB

Hg Diamagnetic

Cd Diamagnetic

Zn Diamagnetic

IIB

Tl Diamagnetic

In Diamagnetic

Ga Diamagnetic

IIIB

Pb Diamagnetic

Sn 0.002

IVB

Source: K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc.

Be Diamagnetic

IIA

Li 0.014

IA

Table 6.15 Magnetic Susceptibilities (×103 ) of Common Paramagnetic Metals at 20◦ C

Bi Diamagnetic

VB

608

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

(see Figure 6.52a). Typically, susceptibilities at 0.9θc are approximately half those at absolute zero. In ferromagnetic materials, the exchange interaction gives rise to a parallel alignment of spins. In some metals like Cr, this exchange interaction, which is extremely sensitive to interatomic spacing and to atomic positions, results in an antiparallel alignment of spins. When the strength of antiparallel spin magnetic moments are equal, no net spin moment exists and the resulting susceptibilities are quite small. Such materials are called antiferromagnetic. The most noticeable characteristic of an antiferromagnetic material is that the susceptibility attains a maximum at a critical temperature, θN , called the N´eel temperature. The maximum in susceptibility as a function of temperature can be explained using thermodynamic reasoning. At absolute zero, the spin moments are aligned as antiparallel as possible, yielding a small susceptibility. With increasing temperature, the tendency for disorder or more random alignment of spin moments become greater,

Table 6.16 Distribution of Electrons and Electronic Magnetic Moments in Some Transition Metals Element

Fe

Co

Ni

Cu

Total number (s + d) valence electrons Average number of electrons in s band Average number of electrons in d band Number of d electrons with moments up Number of d electrons with moments down Net moment per atom (units of Bohr magnetons, µB )

8 0.2 7.8 5.0 2.8 2.2

9 0.7 8.3 5.0 3.3 1.7

10 0.6 9.4 5.0 4.4 0.6

11 1.0 10.0 5.0 5.0 0

Source: K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc.

Very large

c = c

Very large

C T − qD

qC

T

Moments below qC (a)

c =

c

C T+q

c

qN

T

Moments below qN (b)

qN

T

Moments below qN (c)

Figure 6.52 The magnetic spin alignments and variation of susceptibility with temperature for (a) ferromagnetic, (b) antiferromagnetic, and (c) ferrimagnetic materials. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

MAGNETIC PROPERTIES OF MATERIALS

609

and a higher net moment results in the presence of an applied magnetic field. This same process accounts for the decrease in net moment of ferromagnetic materials between absolute zero and θc . For antiferromagnetic materials, the spin moments become essentially random at θN , and above this temperature, paramagnetic behavior with decreasing χ is observed (see Figure 6.52b). Thus, the N´eel temperature of an antiferromagnetic temperature is analogous to the Curie temperature of a ferromagnetic material. We will discuss antiferromagnetic materials in further detail in a later section. Ferrimagnetism is similar to antiferromagnetism in that the spins of different atoms or ions tend to line up antiparallel. In ferrimagnetic materials, however, the spins do not cancel each other out, and a net spin moment exists. Below the N´eel temperature, therefore, ferrimagnetic materials behave very much like ferromagnetic materials and are paramagnetic above θN (see Figure 6.52c). Ferrites, a particularly important class of ferrimagnetic materials, will be described later. The different types of magnetism described so far are summarized in terms of the sign and magnitude of their susceptibilities in Table 6.17. A summary of the terms, symbols, and units used in magnetism are provided in Table 6.18, in a manner analogous to the electrical terms in Table 6.5. 6.2.1.3 Magnetic Domains and Hysteresis. In addition to susceptibility differences, the different types of magnetism can be distinguished by the structure of the magnetic dipoles in regions called magnetic domains. As shown in Figure 6.53, each domain consists of magnetic moments that are aligned, giving rise to a permanent net magnetic moment per domain. The domains are separated by domain boundaries (not the same as the grain boundaries), or domain walls, across which the direction of the spin alignment gradually changes. Typically, the domain wall has a thickness of about 100 nm. This type of arrangement represents the lowest free energy in the absence of an externally applied magnetic field. When the bulk material is unmagnetized, the net magnetization of these domains is zero, because adjacent domains may be oriented randomly in any number of directions, effectively canceling each other out.

Table 6.17 Summary of the Various Types of Magnetism

Type of Magnetic Behavior

Characteristics of Magnetic Susceptibility Typical Materials

Sign

Magnitude

Diamagnetism

Negative

Smalla χ = constant

Organic materials, superconducting metals, and other metals (e.g., Bi)

Paramagnetism

Positive

Small χ = constant

Alkali and transition metals, rare earth elements

Ferromagnetism

Positive

Large χ = f (H )

Some transition metals (Fe, Ni, Co) and rare earth metals (Gd)

Antiferromagnetism

Positive

Ferrimagnetism

Positive

a

Small χ = constant

Large χ = f (H )

Salts of transition elements (MnO) Ferrites (MnFe2 O4 , ZnFe2 O4 ) and chromites

Diamagnetic susceptibilities for superconducting metals are large.

610

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Table 6.18 Magnetic Symbols, Terms, and SI Units SI Units Quantity

Symbol

Primary

Magnetic field strength

H

Magnetic induction (magnetic flux density)

B

C/m · s

kg/s · C

Magnetic susceptibility

χ

(dimensionless)

Derived A/m = 10−3 oersted (Oe)a

V · s/m2 = Weber (Wb)/m2 = tesla (T) = 104 gauss (G)a (dimensionless)

C/m · s

Magnetization

M

Magnetic dipole moment

pm

Remnant induction

Br

kg/s · C

V · s/m2 = weber/m2 = tesla ∗ (T) = 104 gauss(G)

Coercivity (coercive force)

Hc

C/m · s

A/m = 10−3 oersted(Oe)

Magnetic permeability in vacuum

µ0

kg · m/C2

Henry/m = Wb/A · m

Saturation magnetization

Ms

C/m · s

A/m

a

C · m2 /s

A/m A · m2



cgs unit; items in boldface are material properties.

I

One domain

Another domain

Domain wall

Figure 6.53 Schematic illustration of magnetic domains.

The average magnetic induction of a ferromagnetic material is intimately related to the domain structure. When a magnetic field is applied, the domains most nearly parallel to the direction of the applied field grow in size at the expense of the others. This is called boundary displacement of domains or domain growth. This process continues until only the most favorably oriented domains remain. The spontaneous moments are aligned naturally with respect to certain crystallographic directions; in the case of iron, they are parallel to the [100] direction. When domain growth is completed, a further increase in the magnetic field causes the domains to rotate and align parallel

MAGNETIC PROPERTIES OF MATERIALS

611

to the applied field. The material reaches the point of saturation magnetization, and no further change will take place on increasing the strength of the magnetic field. This change in magnetization with applied magnetic field is shown schematically in Figure 6.54, in a diagram known as a hysteresis loop. Hysteresis is defined as the lag in magnetization change that follows variations in the applied magnetic field. If an unmagnetized specimen, such as iron, is placed in a magnetic field, H , a magnetic induction, B, results, as described earlier. When the field is increased, the magnetic induction will also increase, as shown in curve (a) of Figure 6.54. This is the domain growth region. At sufficiently high magnetic fields, all of the magnetic dipoles become aligned with the external field, and the material becomes magnetically saturated; that is, the magnetization becomes constant and is called the saturation magnetization, Ms . There is a corresponding saturation flux density, Bs , as the induction reaches a maximum value. Once magnetic saturation has been achieved, a decrease in the applied field back to zero, as illustrated in curve (b) of Figure 6.54, results in a macroscopically permanent induction, called the remnant induction, Br , under which the material is in the magnetized condition, even at zero applied field. At this point, spin orientations within domains have readily rotated back to their favorable crystallographic positions, but the original random domain arrangement is not achieved because domain wall motion is limited and the domain growth process is not entirely reversible. To bring the induction back to zero, a field of magnitude Hc , termed the coercivity, must be

B (b )

+Br (a) (c ) −H

H −Hc

+Hc (c)

−Br

+H

(a) = Domain growth (b) = Field removal (c) = Hysteric curve on field

Figure 6.54 A hysteresis loop for a ferromagnetic materials. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission John Wiley & Sons, Inc.

612

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

applied antiparallel to the original magnetic field, shown as curve (c) in Figure 6.54. A further increase in the magnetic field in the negative direction results in a maximum induction and saturation magnetization, but in the opposite direction of the previous maximum achieved under an applied field in the forward direction. The field can once again be reversed, and the field-induction loop can be closed. The area within the hysteresis loop represents the energy loss per unit volume of material for one cycle and, unless suitable cooling is provided, the material heats up. Smaller hysteresis loops result for cycling between applied fields that do not cause saturation magnetization.

Cooperative Learning Exercise 6.6

Recall that the saturation magnetization, Ms , is the maximum possible magnetization in the material, and is simply the product of the net magnetic moment per atom, Pm , and the number of atoms per unit volume, N. The net magnetic moment, in turn, is related to the electronic structure (paired or unpaired electrons), although a number of other factors come into play. Use this information to calculate the saturation magnetization for nickel. Person 1: Look up the net magnetic moment of nickel in Table 6.16 and convert it from units of Bohr magnetons to A · m2 . Person 2: Look up the density and atomic weight of nickel in Tables 1.11 and 1.3, respectively, and calculate the number of atoms per cubic meter for nickel. (Don’t forget Avogadro’s number.) Combine your information to calculate Ms for nickel in units of A/m. Since the magnetic susceptibilities of ferromagnetic materials are so high, H > n. If the size of the clusters get so big as to form magnetic domains, the advantage is lost. Ferrite can be used as soft magnets, hard magnets, and a class of materials known as semihard magnets. Soft magnetic ferrite is used for core materials for coils and transformers. Hard magnetic ferrite is used in making permanent magnets. Semihard magnetic ferrites have properties between those of soft and hard magnetic ferrites. They are comparatively easy to magnetize, can retain magnetization, and can produce changes in the magnetic flux from this state, as needed. The typical application for semihard magnetic ferrites is in memory cores for computers. In general, the Mn–Mg ferrites and the lithium ferrites are useful for semihard applications. There is a growing list of applications for ferrites, including microwave applications that utilize the unique magnetic properties of these materials. 6.2.2.3 The Meissner Effect. In Section 6.1.2.4, we describe the interrelationship between electric current density, magnetic field, and temperature for oxide superconductors, and although a great deal of description was given to upper and lower critical magnetic fields, one particularly important topic related to magnetism was deferred to this section. That topic has to do with an important phenomenon related to superconductivity called the Meissner effect. Like superconductivity itself, the Meissner effect is relatively old as far as scientific discoveries go. It was found in 1933 by W. Meissner and colleagues, who determined that a superconducting metal expels any magnetic field when it is cooled below the critical transition temperature, Tc , and becomes superconducting. By expelling the field and thus distorting nearby magnetic field lines, as shown in Figure 6.64b, a superconductor will create a strong enough force field to overcome gravity. This expulsion of the magnetic field is to be distinguished from simply preventing the field lines from penetrating—any metal with infinite conductivity would do the latter. If a magnetic field is already present, and a substance is cooled below Tc to become a superconductor, the magnetic field is expelled. The significance of this difference cannot be explained merely by infinite conductivity. The ability of the superconductor to overcome gravitational forces has led to the famous demonstration, typically involving an oxide superconductor and a rare-earth magnet, in which the magnet is levitated above the superconductor when it is cooled

626

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

A

(a) NORMAL

A′

(b) SUPERCONDUCTING

Figure 6.64 Schematic illustration of Meissner effect in which magnetic flux lines that (a) normally penetrate the material are (b) expelled in the superconducting state.

with liquid nitrogen. In this experiment, the magnetic field lines from the strong (but small) magnet penetrate the superconductor to a small depth called the penetration depth, λ, which sets up circulating currents—called eddy currents —in the superconductor. Typically, the penetration depth is less than 0.05 µm, so that on a macroscopic scale, the superconductor is said to exclude the magnetic field. In actuality, these eddy currents create their own magnetic field lines that exactly cancel the applied magnetic field from the magnet, thus levitating it. No superconductor can keep out very strong magnetic fields. As described previously, Type I superconductors have a critical magnetic field, Hc , above which they no longer superconduct. Type II superconductors have a range of critical fields from the upper critical field, Hc2 , to the lower critical field, Hc1 . Below these critical fields, the superconductor is perfectly diamagnetic. For a perfect diamagnet, M = −H , so that according to Eq. (6.61) the inductance is B = 0. Thus, the earlier statement that diamagnetism is not important in ceramic materials must be strongly qualified for the oxide superconductors—diamagnetism is critical to their importance. At first, the superconducting state was not thought to be a thermodynamic equilibrium state. But, as we know from Chapter 2, any equilibrium state must be the result of a free energy minimization. It can be shown that the superconducting and normal states will be in equilibrium when their free energies are equal, and that the free energy difference between the normal state at zero field, Gn (H = 0), and the superconducting state at zero field, Gs (H = 0), is Gn (H = 0) − Gs (H = 0) =

µ0 Hc2 2

(6.66)

The free energies for the respective states can be determined from heat capacity data, and the thermodynamic critical field can be determined from a plot of free energy versus applied field for the two states, as illustrated in Figure 6.65. The Meissner effect is a very important characteristic of superconductors. Among the consequences of its linkage to the free energy are the following: (a) The superconducting state is more ordered than the normal state; (b) only a small fraction of the electrons in a solid need participate in superconductivity; (c) the phase transition must be of second order; that is, there is no latent heat of transition in the absence of any applied magnetic field; and (d) superconductivity involves excitations across an energy gap.

MAGNETIC PROPERTIES OF MATERIALS

627

Gs

Gn

Su

pe

rc o

nd

uc tin

g

St at e

Free Energy Density

Normal State

Hc Applied Magnetic Field H

Figure 6.65

6.2.3

Free energy curves for both normal and superconducting states.

Magnetic Properties of Polymers

Much like ceramics, the electrical and magnetic properties of polymers are closely linked. Initially, this statement can be viewed with considerable disappointment, since, with only a few notable exceptions, polymers are poor electrical conductors, and what is known about their potential for magnetization is even less than is known about their electrical conductivity. Nonetheless, there is hope. Just as new classes of polymers are currently being developed with improved electrical conductivities, there is reason to believe that polymers can be developed that will have significant magnetic properties. What a potential benefit this could be for the development of lightweight magnets. We briefly describe here just one topic related to magnetic polymers that is still too “young” in its scientific development to warrant significant coverage in an introductory text: molecular magnets. The more common use of polymers in magnetic media, that as a matrix for the support of nonpolymeric magnetic particles, is more appropriately addressed in the section on composites. 6.2.3.1 Molecular Magnets∗ . Much of the information presented here on molecular magnets is drawn from some excellent, recent review articles [9–12] on this emerging field. Following a brief introduction to the general topic of molecular magnets, a description of organic molecular magnets will be given, with an emphasis on the development of polymeric molecular magnets. A molecular magnet is a magnet that is molecule-based—that is, composed of molecular or molecular ion entities, and not atoms or atomic ions. This distinction is not new to us, though previously we have not elaborated upon it. Simply compare the crystal structures of such atom-based solids as BaTiO3 in Figure 1.42, in which the

628

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

positions of the atoms determine the crystal structure, with that of a molecule-based solid such as polyethylene in Figure 1.65, in which the positions of the molecules, in this case chains, determine the crystal structure. Think about this distinction for a moment, and you will realize that most molecular materials are organic in nature. The covalent bonds within the molecules are stronger than the noncovalent bonds that draw the organic compounds into a molecular crystal, such as hydrogen bonding and van der Waals interactions. This is different from metallic and ionic solids, in which any Na–Cl interaction, for example, is essentially the same anywhere in the crystal, so that ordering takes place on the atomic scale. The magnetic materials described so far in this chapter are all atom-based: Their bulk magnetic properties are due to the incorporation of atoms such as Fe or Nd. In these materials, magnetic order arises from cooperative spin–spin interactions between unpaired electrons located in d or f orbitals of the metal atoms. Molecular-based materials, on the other hand, do not tend to exhibit magnetic properties for some very good reasons. Organic compounds generally have closed-shell structures and are thus diamagnetic. Paramagnetic organic molecules are very rare because the presence of unpaired electrons—which are required for paramagnetism—in the s and p orbitals gives instability to the molecules, which tend to spontaneously form a covalent bond with another molecule, thus pairing the spins. Even if one or more unpaired electrons are maintained stably in an organic molecule, stabilization of a triple state (parallel alignment of spins) requires that certain quantum mechanical conditions be met which are difficult to satisfy, such that they prefer to couple in an antiferromagnetic manner with no spontaneous magnetic moment. Thus, the development of ferro- or ferrimagnetic organic molecules has been elusive. However, recent advances in synthetic molecular chemistry have made possible the formation of organic molecules with sufficient magnetic function that can serve as building blocks for bulk magnetic behavior. Without delving into the theory of magnetism in these materials, we describe some of these building blocks that may one day be the precursors to polymer magnets. The typical synthetic approach to designing molecular magnets consists of choosing molecular precursors, each of which contains an unpaired spin, as represented by the arrow in Figure 6.66. The precursors are then assembled in such a way that there is

Functionalized Molecule as Building block

Figure 6.66 Schematic illustration of building blocks used to construct molecular magnets. An arrow indicates an unpaired electron, and a star represents some other desirable functionality. Reprinted, by permission, from J. V. Yakhmi, Physica B, 321, 206. Copyright  2002 by Elsevier Science B. V.

MAGNETIC PROPERTIES OF MATERIALS

629

no compensation of the spins at the crystal lattice scale. In Figure 6.66, the precursors may contain other functional groups, represented by the star, that can assist in selfassembly, or impart some additional property such as chirality or ferroelectricity. The primary criteria for designing a molecular magnet, then, are that (a) all the molecules in the lattice have unpaired electrons, and (b) the unpaired electrons have spins aligned parallel along a given direction. Since magnetism is a cooperative effect, the spin–spin interaction must extend to all three dimensions in the lattice. These criteria are met in molecules such as the nitronyl nitroxide radical, NITR, shown in Figure 6.67a, where R is an alkyl group. Through different substitutions for the alkyl group, the spin delocalization effect can be fine-tuned. An example is when R is a nitrophenyl compound, which results in 4-nitrophenyl nitronyl nitroxide, or p-NPNN for short, as shown in Figure 6.67b. Ferromagnetism has been exhibited in this purely organic compound at around 0.6 K. Other similar, organic compounds have Curie temperatures of 16 K and above. An alternative strategy is to utilize ferrimagnetic chains containing alternating spins of unequal magnitude and assemble them in such a way that there is a net spin. For example, two different spin carriers such as Mn ions, with ms = 5/2, and Cu ions, with ms = 1/2, can be incorporated within the same molecular precursor, as schematically illustrated in Figure 6.68. It is also possible to create different spins using organic radicals instead of metal ions. An example of such a building block is shown in Figure 6.69 for [Cu(opba)]2− , where “opba” stands for ortho-phenylenebis(oxamato). This copper

O− N +

O− N +

R

(a)

NO2

N

N O∗

(b)

O∗

Figure 6.67 Chemical structures of (a) NITR and (b) p-NPNN used in molecular magnets. An asterisk, *, indicates an unpaired electron. Reprinted, by permission, from J. V. Yakhmi, Physica B, 321, 206. Copyright  2002 by Elsevier Science B. V.

Cu

Cu Mn

Cu

Cu Mn

Cu Mn

Cu Mn

Figure 6.68 Schematic illustration of spin assembly metal-based ferrimagnetic chains. Reprinted, by permission, from J. V. Yakhmi, Physica B, 321, 206. Copyright  2002 by Elsevier Science B. V.

630

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

O

O

N

O

2−

Cu N

O

O

O

Figure 6.69 Structure of the copper dianion precursor [Cu(opba)]2− . Reprinted, by permission, from J. V. Yakhmi, Physica B, 321, 206. Copyright  2002 by Elsevier Science B. V.

precursor can be reacted with a divalent ion such as Mn2+ , which results in an amorphous magnet with a spontaneous magnetization below 6.5 K. Similar precursors can be assembled into crosslinked chains with still higher Curie temperatures, as shown in Figure 6.70. Note that even though these structures contain transition metals such as Mn and Cu, it is the entire molecule, including the metal ion, that exhibits the spin, and that can be used as a building block for larger structures. One final example is a class of related materials that undergo reversible magnetic behavior through a mild dehydration-rehydration process. Because these materials show a reversible crossover under dehydration to a polymerized long-range, magnetically ordered state with spontaneous magnetization, and transformation back to isolated molecular units under rehydration, they are called molecular magnetic sponges. Coercivity values for these sponges are high (see Table 6.23), and some even undergo color changes that coincide with the change in magnetic properties. In addition to the “opba” anion described above, these molecular magnetic sponges are composed of “obbz” and “obze” ligands, where obbz = oxamido bis(benzoato) and obze = oxamido-N benzoato-N ′ -ethanoato. As this field continues to develop, look for improvements not only in the magnetic properties of these organic polymers, but also in the theoretical descriptions of intrinsic ferro- and ferrimagnetism in polymers. 6.2.4

Magnetic Properties of Composites

In this section, we describe the properties of two magnetic composites that represent extremes in terms of applications for magnetic materials. The first, flexible magnets illustrate how the incorporation of less-expensive components into a material can retain the functionality of the magnetic material while significantly reducing cost. The second, magnetic storage media illustrates how a simple concept such as a reversible magnetic moment can be utilized to create an entirely new industry and revolutionize the manner in which we conduct our daily lives. 6.2.4.1 Flexible Magnets. Flexible magnets are made by embedding ferrimagnetic ceramics in a polymer matrix, such as barium ferrite or strontium ferrite in nylon or

MAGNETIC PROPERTIES OF MATERIALS

631

a

b

c

Figure 6.70 Structure of 1:2 Mn(II)Cu(II) compound consisting of a zigzag chain with terminal “opba” groups. Reprinted, by permission, from J. V. Yakhmi, Physica B, 321, 207. Copyright  2002 by Elsevier Science B. V.

Table 6.23 Selected Magnetic Properties of Some CoCu-Based Molecular Magnetic Sponges Compound CoCu(pbaOH)(H2 O)3 · 2H2 O CoCu(pba)(H2 O)3 · 2H2 O CoCu(obbz)(H2 O)4 · H2 O CoCu(obze)(H2 O)4 · 2H2 O

Tc (K)

Hc (kOe)

38 33 25 20

5.66 3 1.3 1

Source: J. V. Yakhmi, Physica B, Vol. 321, pp. 204–212. Copyright  2002 by Elsevier Science B.V.

polyphenylene sulfide. In an anisotropic magnet (one-sided) the ferrite and polymer components are mixed and processed, usually by extrusion, into a thin sheet and are then coated or printed on one side. Typical tape thicknesses are 1–2 mm. To obtain maximum alignment, the ferrite particles are physically rotated in a magnetic field during the molding process. In this way, energy products equivalent to those of cobalt steel are obtained.

632

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Side view field lines

back (brown) side

front (printed) side

Side view

domains (arrows indicate field direction) field lines

Figure 6.71 Side view of magnetic flux lines in an anisotropic flexible magnet. (mrsec.wisc.edu/edetc).

The magnetic field structure (see Figure 6.71) in an anisotropic magnet is made of a striped pattern of alternating pole alignments on the back side of the flexible magnetic sheet. This arrangement also channels the magnetic field to the back of the magnet so that maximum holding power is obtained, which is important for many applications. It is also possible to make isotropic, or two-sided, magnetic tapes and sheets, for which the field lines shown in Figure 6.71 are present on both sides of the sheet. The magnetic properties of the flexible magnets depend upon the properties of the ferrite, of course, and the relative amounts of ferrite and binder, as well as processing conditions, but generally for strontium ferrite-based materials, residual inductions, Br , range from 1.6 to 2.6 gauss, coercive forces, Hc , range from 1.3 to 2.3 kOe, and densities range from 3.5 to 3.8 g/cm3 . The magnetic properties degrade linearly with increasing temperature, and there is generally an upper-use temperature of around 100◦ C at which point the polymeric component begins to melt or flow. 6.2.4.2 Magnetic Storage Media. Magnetic materials are ideal candidates for information storage media because magnets can be used like switches. They can be made to point in one direction or another in the presence of a magnetic field. Reverse the direction of the field, and the magnet will switch to point in the other direction. This represents a simple way to store binary information. The magnet represents a “1” when pointing in one direction and represents a “0” when pointing in the other. A magnetic memory can be constructed from an entire array of magnets, in which the binary data can be encoded by applying an external magnetic field to each magnet in turn, the direction of which determines whether a “1” or a “0” is written. As we have seen, individual atoms can act as magnets. In a ferromagnetic material such as iron, for example, the magnetic moments in each atom tend to line up so that they all point in the same direction. Furthermore, recall that the magnetic moments tend to align in domains, separated by domain walls, the net magnetic moments of which are randomly oriented to create a bulk material with no net magnetic moment.

MAGNETIC PROPERTIES OF MATERIALS

633

The ability to influence the alignment of magnetic moment by the application of a magnetic field forms the basis of magnetic recording. By applying a strong, localized magnetic field to one region of a thin layer of a ferromagnetic material such as magnetite (above the Curie temperature), a domain can be created in which all the magnetic moments are aligned. The magnetized domains, in general, are larger than the magnetic particles. Application of the field in the opposite direction in another part of the magnetic medium creates a domain of opposite alignment. One alignment direction represents the “1” in the binary code, and the opposite direction represents the “0.” The alignment is generated by a recording head, which generates a magnetic field that changes the direction of alignment in the magnetic particles, as shown in Figure 6.72. The recording head contains an electromagnet in which the field is generated by flowing current across a small gap. When the recording head is magnetized, a magnetic field bridges the gap, and the flux lines spread out so that they pass through the magnetic medium lying below. The magnetic moments in the medium align themselves with the direction of the field across the recording head gap. Essentially the same process in reverse is used to retrieve information from the device. Magnetic tape technology, while still important in limited applications, has been almost completely replaced by two-dimensional disk technology due to the limitations of the one-dimensionality of the tape. In order to retrieve information from one part of the tape, the entire preceding information must first be trawled through. Though the disk allows information to be retrieved from a specific site on the disk surface, the principles of magnetization and storage of binary information via alignment of magnetic particles is essentially the same. The primary principle leading to the development of new materials in this area is storage density. As illustrated in Figure 6.73, there must be a transition region between each oriented domain in the magnetic film. This buffer zone allows adjacent magnetic domains to retain their alignment without influencing each other. The resistance to alignment flipping due to an adjacent domain is related to the coercivity of the particles. The higher the coercivity, the more resistant the particle

Electric current

Core Demagnetizing field

Magnetization

Write head Gap

Tape motion

Magnetic medium

Substrate

Figure 6.72 Schematic diagram of a magnetic recording device and magnetization in magnetic tape recording medium. Reprinted, by permission, from P. Ball, Made to Measure, p. 72. Copyright  1997 by Princeton University Press.

634

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Demagnetizing field

−− −− −− −−

++ ++ ++ ++

Magnetized region

−− −− −− −−

++ ++ ++ ++

−− −− −− −−

Transition region

Figure 6.73 Schematic illustration of transition region between magnetic domains in a magnetic information storage medium. Reprinted, by permission, from P. Ball, Made to Measure, p. 75. Copyright  1997 by Princeton University Press.

is to flipping, and the smaller the transition zone between domains can be. Smaller transition zone and magnetic domains translate directly into increased storage density. Composite magnetic recording media consist of submicroscopic, single-domain particles of magnetic oxides or metals immersed in a polymeric binder that separates the particles and binds them to the substrate. The substrates used in tapes, magnetic cards, and floppy disks are generally made from poly(ethylene terephthalate), while rigid disks are fabricated from an Al–Mg alloy. Particles of γ -Fe2 O3 have been used in tapes for a long time, but as the bit length of recorded signals becomes shorter, further improvements in coercivity are required. Coercive fields have been raised from 100 to 500 A/m by impregnating the surface of the iron oxide particles with cobalt. In addition to magnetic particle/polymer composites, laminar magnetic composite thin films are used for information storage. The magnetic film is typically either a CoPtCr or CoCrTa alloy of thickness 10 to 50 nm on a substrate of pure chromium or chromium alloy (see Figure 6.74). The carbon layer overcoat is to improve mechanical stability, reduce corrosion of the magnetic layer, and provide a hard, low-friction surface for the head to glide over. The magnetic thin film is polycrystalline, with an average grain size of 10–30 nm. Each grain is a single magnetic domain, with the direction of easy magnetization for each grain aligned in the direction of disk motion. In these materials, it is desirable to have a relatively large and square hysteresis loop. Saturation flux densities range from 0.4 to 0.6 tesla, and coercivities range from

Lubricant, ∼1 nm Carbon overcoat, 1)

Scattered light ray

Figure 6.94 Schematic illustration of light scattering in a solid due to a pore. Reprinted, by permission, from J. F. Shackelford, Introduction to Materials Science for Engineers, 5th ed., p. 599. Copyright  2000 by Prentice-Hall, Inc.

scattering result can be created with second-phase particles such as SnO2 with a higher index of refraction (n = 2.0) than that of the medium, such as glass (n = 1.5). Such materials are called opacifiers. The degree of scattering and opacification created by the pores or particles depends on their average size and concentration, as well as the mismatch between the refractive indices. If individual pores or particles are significantly smaller than the wavelength of light, they are ineffective scattering centers. Conversely, particles or pores in the 400- to 700-nm size range maximize the scattering effect. This effect is what leads to transparent ceramics, such as glass ceramics (cf. Section 1.2.5) where the crystallite sizes are carefully controlled through nucleation and growth to be less than the wavelength of visible light (cf. Figure 1.52). 6.3.2.5 Decay Processes: Luminescence and Lasers∗ . The absorption and reemission of electromagnetic radiation leads to some interesting phenomena that can be utilized for important applications. Both lasers and luminescence are built upon the principle of emission processes and will be described here in brief. As we have already seen, the absorption of photons by a material can lead to the reemission of other photons of equal or lesser energy. When photon absorption is accompanied by emission of photons in the visible spectrum, the process is called photoluminescence, or simply luminescence. In fact, luminescence can be used to describe the emission of visible light as a result of absorption of any number of forms of energy, including thermal, mechanical, or chemical energy. The photon emission is the result of excited electrons returning to their ground state. The duration of this relaxation process is used to distinguish between the two main types of luminescence: fluorescence and phosphorescence. If the relaxation process occurs in less than about 10 ns (10−8 s), it is termed fluorescence. The fluorescent process also involves additional internal or

660

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Singlet excited states Internal conversion

S2

Triplet excited state Vibrational relaxation Intersystem crossing

S1

Energy

T1

Absorption

Fluorescence

Phosphorescence

Vibrational relaxation

S0 Ground state

Internal and external conversion

l2

l1

l3

l4

Figure 6.95 Schematic diagram of transitions between electronic energy levels involved in fluorescence and phosphorescence. Reprinted, by permission, from D. A. Skoog and D. M. West, Principles of Instrumental Analysis, 2nd ed., p. 282. Copyright  1980 by Saunders College.

external energy conversion, as illustrated with the energy diagrams in Figure 6.95. In this way, the energy of the reemitted photons is less than the energy of the absorbed photons. If the reemission process takes longer than about 10 nanoseconds, it is termed phosphorescence. Luminescence is generally produced through the addition of impurities to compounds. For example, zinc sulfide containing small amounts of excess zinc, or impurities such as gold or manganese, emits light in the visible spectrum after excitation with ultraviolet radiation (see Figure 6.96). Zinc sulfide and a similar compound, Mn-activated zinc orthosilicate, Zn2 SiO4 , are also examples of compounds used as phosphors. Phosphors are useful for detecting X-ray or γ -ray radiation because they emit visible light upon exposure. They are used for other, more mundane purposes as well, such as the luminous hands on watches and clocks. Light amplification by stimulated emission of radiation, or LASER, is another application of radiative relaxation processes. Unlike most radiative processes, such as luminescence, which produce incoherent light (light waves are out of phase with each other), the light produced by laser emission is coherent (light waves are all in phase with each other). The principal benefit of coherent light is that the photons do not disperse as they do in an incoherent beam of light. Consequently, the laser beam does not spread out in the same way that an incoherent beam can. Additionally, lasers produce a beam of light that is monochromatic—that is, entirely of one wavelength. This is

OPTICAL PROPERTIES OF MATERIALS

661

Emission intensity

Cu

0.4

Mn Ag Zn

0.5 l (µm)

0.6

Figure 6.96 Luminescence spectra for Zn doped with Cu, Ag, Mn, and excess Zn. From W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

in contrast to most light sources, even phosphors, that produce a range of wavelengths, or polychromatic light. Polychromatic light is not entirely a bad thing—we much prefer white light (all or most of the wavelengths in the visible spectrum) to monochromatic light for our daily and nightly operations. But the monochromatic, coherent, and collimated (photons directed along the same straight line) beam of a laser opens up a realm of possible applications. Let us see how this type of beam can be produced. The radiative transitions of the previous descriptions have all been spontaneous: Relaxation from the excited state to the ground state and emission of photons occur without external aid. In contrast, a stimulated emission occurs when the halflife of the excited state is relatively long, and relaxation can occur only through the aid of a stimulating photon. In stimulated emission, the emitted photon has the same direction as, and is in phase with, the stimulating photon. The example of Cr3+ doped Al2 O3 that we utilized earlier for our description of the color of ruby works equally well for a description of stimulated emission. Recall that the presence of chromium in alumina alters the electronic structure, creating a metastable state between the valence and conduction bands. Absorption of a blue-violet photon results in the excitation of an electron from

HISTORICAL HIGHLIGHT

The physical principles that give rise to the emission of coherent light were elucidated in the first half of [the 20th] century, by Albert Einstein among others. But not until 1954 were these principles put into practice in a working device that emitted coherent radiation. This device, developed by Charles Townes of the University of California at Berkeley, was not really a laser at all but a maser, which generated coherent microwave radiation rather than light. Townes and others strove in the ensuing years to create devices that would emit coherent light at visible wavelengths, and the first genuine laser was demonstrated in 1960 by H. Maiman, who obtained red laser light from a ruby laser. Source: Ball, Made to Measure, p. 29.

662

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

the ground state to the excited state, as illustrated in Figure 6.97. The excited electron relaxes to a metastable electronic state through a non-radiative process involving the transfer of energy to a phonon. Spontaneous decay results in the emission of a photon with wavelength in the red part of the visible spectrum (λ = 694.3 nm). However, most of the electrons reside in this metastable state for up to 3 ms before decaying to the ground state. This time span is sufficiently long that several electrons can occupy this metastable state simultaneously. This fact is utilized to create artificially large numbers of electrons in the metastable state through a process known as optical pumping, in which a xenon flash lamp excites a large number of electrons into the excited state, which decay rapidly (in the nonradiative process) to the metastable state. When spontaneous emission finally occurs due to a few electrons, these electrons stimulate the emission of the remaining electrons in the metastable state, and an avalanche of emission occurs. The photons are of the same energy and are in phase (coherent). The beam is collimated through the use of a tube with silvered mirrors at each end, as illustrated in Figure 6.98. As stimulated emission occurs, only those photons traveling nearly parallel to the long axis of the ruby crystal are reflected from the fully silvered surface at the end of the tube. As these reflected photons travel back through the tube, they stimulate the emission of more photons. A fraction of these photons are re-reflected at the partially silvered face at the other end of the tube. Thus, photons travel up and down the length of the crystal, producing ever-increasing numbers of stimulated photons. When the beam is finally emitted through the partially silvered face, it is a high-energy, highly collimated, monochromatic beam of coherent light. Typical power densities for a ruby laser are on the order of a few watts. In addition to ruby, lasers can be constructed from gases such as CO2 or He–Ne and from semiconductors such as GaAs (cf. Section 6.1.2.5) and InGaAsP. Gas lasers generally produce lower intensities and powers, but are more suitable for continuous operation since solid-state lasers generate appreciable amounts of heat. The power in a solid-state laser can range from microwatts to 25 kW, and more than a megawatt in lasers for defense applications. Some common lasers and their associated powers are listed in Table 6.27. Lasers that can selectively produce coherent light of more than one wavelength are called tunable lasers.

Excited states

Metastable state

Metastable Excitation

l = 694.3 nm Emission Ground state

Figure 6.97 Schematic illustration of energy levels in ruby that are used to create a populated metastable electronic state, which can then be stimulated to emit monochromatic, coherent radiation for a laser. Reprinted, by permission, from J. F. Shackelford, Introduction to Materials Science for Engineers, 5th ed., p. 607. Copyright  2000 by Prentice-Hall, Inc.

OPTICAL PROPERTIES OF MATERIALS

663

Flash lamp Ruby Coherent beam

Partially silvered face

Fully silvered face

Power source

Figure 6.98 Schematic illustration of a ruby laser. From K. M. Ralls, T. H. Courtney, and J. Wulff, Introduction to Materials Science and Engineering. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc. Table 6.27 Characteristics and Applications of Some Common Lasers

Laser

Type

Common Wavelengths (µm)

Maximum Output Power (W)a

Applications

He-Ne

Gas

0.6328, 1.15, 3.39

0.0005–0.05 (CW)

Line-of sight communications, recording/playback of holograms

CO2

Gas

9.6, 10.6

500–15,000 (CW)

Heat treating, welding, cutting, scribing, marking

Argon

Gas ion

0.488, 0.5145

0.005–20 (CW)

Surgery, distance measurements, holography

HeCd

Metal vapor

0.441, 0.325

0.05–0.1

Light shows, spectroscopy

Dye

Liquid

0.38–1.0

Spectroscopy, pollution detection

Ruby

Solid state

0.694

0.01 (CW) 1 × 106 (P) (P)

Pulsed holography, hole piercing

Nd-YAG

Solid state

1.06

Welding, hole piercing, cutting

Nd-glass

Solid state

1.06

1000 (CW) 2 × 108 (P)

5 × 1014 (P)

Pulse welding, hole piercing

Diode

Semiconductor

0.33–40

0.6 (CW) 100 (P)

Bar-code reading, CDs and video disks, optical communications

a

“CW” denotes continuous; “P” denotes pulsed. Source: W. Callister, Materials Science and Engineering: An Introduction, 5th ed. Copyright  2000 by John Wiley & Sons, Inc.

664

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Semiconducting lasers have become increasingly important in information storage, such as in the use of compact discs (CDs) and digital video discs (DVDs). In these materials, a voltage is applied across a layered semiconductor (see Figure 6.99), which excites electrons into the conduction band, just as in the ruby laser. However, the holes created in the valence band play an important role in the emission process. It is the recombination of the electron–hole pairs that results in the emission of photons with a specific wavelength. As in the ruby laser, the ends of the semiconducting “sandwich” are fully and partially silvered, and a voltage is continuously applied to ensure that there is a steady source of holes and electrons. These devices are extremely temperaturesensitive, since the current required to cause lasing, the wavelength of emitted radiation, and the lifetime of the diode all depend on temperature. The conventional III–V compound semiconductors such as GaAs are limited to infrared or red emitting diode lasers and are not capable of generating shorter wavelength light, which is a requirement for size reduction in information storage devices. Short-wavelength LEDs and laser diodes require a different class of semiconductor materials; principally, the group III nitrides. The growth of high-quality (AlGaIn)N single-crystalline layer sequences has been a challenge, and only recent developments in fabrication processes have produced blue lasers that can be used in commercially available devices such as DVD players. The wavelengths of some common compound semiconductors are listed in Table 6.28. 6.3.2.6 Photoconductivity∗ . In Section 6.3.1.2 we saw how electricity can be generated from the surface of a metal when it is bombarded with photons. A similar process occurs when certain semiconductors absorb photons and create electron–hole pairs that can be used to generate current. This process is called photoconductivity and is slightly different from the photoelectric effect, since it is an electron–hole pair that is being generated instead of a free electron and since the energy of the electron–hole pair is related to the band gap energy, and not the energy of the Fermi level. When a semiconductor is bombarded with photons equal to or greater in energy than the band gap, an electron–hole pair is formed. The current that results is a direct function of the incident light intensity. Photoconductive devices consist of a p–n junction called a photodiode, or a p–i–n junction commonly used in a photodetector, an

100 µm

Metal electrode 10 µm

1 µm p+ GaAs

Protonbombarded region

p GaAlAs GaAlAs active region

Emitting area

n GaAlAs n GaAs substrate 200 µm

Figure 6.99 Schematic diagram of a layered compound semiconductor laser. (www.mtmi.vu. lt/pfk/funkc dariniai/diod/led.htm).

OPTICAL PROPERTIES OF MATERIALS

665

Table 6.28 Some Common Compound Semiconductor Lasers and Their Characteristic Wavelengths Compound ZnS ZnO GaN ZnSe CdS ZnTe GaSe CdSe CdTe

Wavelength (µm)

Compound

0.33 0.37 0.40 0.46 0.49 0.53 0.59 0.675 0.785

Wavelength (µm)

GaAs InP GaSb InAs Te PbS InSb PbTe PbSe

0.84–0 0.91 1.55 3.1 3.72 4.3 5.2 6.5 8.5

Source: www.mtmi.vu.lt/pfk/funkc dariniai/diod/led.htm

Semiconductor n-doped semiconductor

Light



+



+

p-doped semiconductor Electric field

Space charge region

Space charge region

Distance

Figure 6.100 Schematic illustration of p-i-n junction used in a photodiode. Reprinted, by permission, from P. Ball, Made to Measure, p. 46. Copyright  1997 by Princeton University Press.

example of which is schematically illustrated in Figure 6.100. When light is absorbed in the undoped region (i region), the electron–hole pair is created, and the two charge carriers are pulled in opposite directions by the applied electric field to the two poles, thus creating current. Silicon can be used to detect infrared wavelength light, and InGaAs is used for longer wavelengths, including visible light. However, optical detectors tend to be relatively slow and have low sensitivity. CdS is widely used for the detection

666

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

of visible light, as in photographic light meters. Photoconductivity is also the underlying principle of the photovoltaic cell, commonly called the solar cell, used for the conversion of solar energy into electricity. 6.3.2.7 Optical Fiber∗ . One of the benefits that lasers have provided is a method for transmitting information in binary code (cf. Section 6.1.1.6). This is easily performed by assigning a “one” to a high-intensity pulse (on) and assigning a “zero” to a low intensity pulse (off). The key to utilizing this technology for telecommunications, however, is the ability to transmit this digital information over long distances. One method for transmitting optical information has been through the optical analog to copper wire, the optical fiber. The conversion from the transmission of information via analog signals (such as telephone conversations) on copper wires to the transmission of information via digital signals on fiber-optic cable has been a rapid one and is due primarily to the advantages of optical fiber over copper wire, some of which are listed in Table 6.29. In Table 6.29, the number of two-way transmissions is a measure of information capacity, and the repeater distance is the distance between amplifiers, indicative of the attenuation loss associated with each medium. It is ironic that even with the proliferation of wireless communications (e.g., cellular telephones), the use of land-based transmission media, such as optical fiber, has grown in a concurrent manner, since for every cellular transmission tower, there needs to be the associated land-based components that connect the tower to the communications “backbone.” Hence, it is worthwhile to describe how this important materials technology works. Optical fiber operates on the principle of total internal reflectance. Recall from Eq. (6.85) that when light encounters a change in refractive index, it is bent at an angle. Consider now the case where light goes through three media, each of differing refractive index, as illustrated in Figure 6.101. The angles of refraction are related to the refractive index change at each of the interfaces, through the more general form of Eq. (6.85) known as Snell’s Law :

and

sin θ1 n2 = n1 sin θ2

(6.90)

n3 sin θ2 = n2 sin θ3

(6.91)

Note that Eq. (6.90) reduces to Eq. (6.85) for the case where medium 1 is air (n1 = 1). Table 6.29 Comparison of Transmission Characteristics for Copper Wire and Optical Fiber

Characteristic

Copper Wire

Diameter, mm Density, g/cm3 Repeater distance, km No. of two-way transmissions Corrosion resistance Susceptibility to EM interference

1.5 8.96 2 24 Low High

Optical Fiber 0.025 2–4 30 1344 High Low

OPTICAL PROPERTIES OF MATERIALS

Incident wave

q1

667

n1

Surface

n2

q2

Surface q3

Primary transmitted wave n3

Figure 6.101 Schematic illustration of light propagating through three media of differing refractive indices.

At some angle of θ2 , the transmitted wave will travel parallel to the second interface (i.e., θ3 = 90◦ ), and all of the light is reflected back into medium 2. This value of θ2 is called the critical angle for total internal reflectance, θc . For example, if medium 3 were air (n3 = 1) and medium 2 were a glass of n2 = 1.5, the condition for which θ3 = 90◦ could be found from Eq. (6.91) to be θc = 42◦ . You should be able to visualize, with the aid of Figure 6.101, that total internal reflectance can also be accomplished by selecting an index of refraction for medium 3 such that θ3 = 90◦ . This is called the core and cladding structure, where medium 1 is air at the entrance to the optical fiber, medium 2 is the core, and medium 3 is the cladding. The core–cladding structure is the principle behind fiber optics, and simply consists of a center core, through which the optical signal is transmitted, surrounded by a cladding of different refractive index. The indices of refraction are selected such that ncladding < ncore . Once the light enters the core from the source, it is reflected internally and propagates along the length of the fiber. Typically, both the core and cladding are made of special types of glass with carefully controlled indices of refraction. A polymeric coating is usually applied to the outside of the fiber for protection only. The sharp change of refractive index between the core and cladding results in what is called a step-index optical fiber, as illustrated in Figure 6.102a. It is also possible, through chemical manipulation of the glass components, to vary the index of refraction in a continuous, parabolic manner, resulting in a graded-index optical fiber, as shown in Figure 6.102b. This results in a helical path for the light rays, as opposed to a zigzag path in a step-index fiber. The digital pulse is less distorted in a graded-index fiber, which allows for a higher density of information transmission. Both step- and graded-index fibers are termed multimode fibers. A third type of optical fiber is called a single-mode fiber, as shown in Figure 6.102c, in which light travels largely parallel to the fiber axis with little distortion of the digital light pulse. These fibers are used for long transmission lines, in which all of the incoming signals are combined, or multiplexed, into a single signal. Core and cladding materials are selected not only on the basis of their refractive index, but also for their processability, attenuation loss, mechanical properties, and dispersion properties. However, density, ρ, and refractive index, n, are critical and have been correlated for over 200 optical glasses, for which the following formula

668

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Intensity

Intensity

Index of refraction,n

Time

Time

Input pulse

Output pulse

Intensity

Intensity

(a)

Time

Time

Input pulse

Output pulse Intensity

Intensity

(b)

Time

Time

Input pulse

(c)

Output pulse

Figure 6.102 Schematic illustration of (a) step-index, (b) graded-index, and (c) single-mode optical fibers. Reprinted, by permission, from J. F. Shackelford, Introduction to Materials Science for Engineers, 5th ed., p. 611. Copyright  2000 by Prentice-Hall, Inc.

applies with an accuracy of 2% or better: n=

ρ + 10.4 8.6

(6.92)

Refractive index can also be found from the weighted average of the refractive index of its components, to a first approximation. High-purity silica-based glasses are used as the fiber material, with fiber diameters ranging between about 5 and 100 µm. The fibers are carefully fabricated to be virtually free of flaws and, as a result, are extremely strong and flexible. We will examine this unique fabrication process in more detail in the next chapter. 6.3.3

Optical Properties of Polymers

As with virtually every other material property, if a polymer possesses optical properties comparable to those of a glass or metal, it will be utilized simply on the basis of weight savings, all other factors being equal. As a result, much of the development in the optical properties of polymers is related to creating glass-like transmissivities or metallike reflectivities. There are also a few unique optical characteristics of polymers that

OPTICAL PROPERTIES OF MATERIALS

669

make them useful for applications for which there is no true glass or metal analogue. We investigate some of these optical characteristics of polymers in this section. 6.3.3.1 Polymer Structure, Transmittance and Birefringence. There is a great deal of variability in the light transmission properties of polymers. Some of this variability is due to structural differences, and some is due to the presence of fillers and colorants that are purposely added to make otherwise transluscent polymers appear opaque. We will focus for the moment on structural differences in polymers that lead to differences in the transmission of light. The primary structural characteristic affecting the transmission of light in polymers is crystallinity. As in ceramic materials, the presence of grain boundaries, pores and particles that are of the order in size of the wavelength of light will create scattering. Crystallites in polymers act as these scattering centers. An example is found in polyethylene. Low-density polyethylene is semicrystalline, with only a few crystallites to act as scattering centers, so that it is mostly transparent. However, high-density polyethylene, which is highly crystalline, is translucent. In fact, most polymer crystals are of the right size to cause scattering, so a rule of thumb is that crystalline polymers are translucent (or opaque), whereas noncrystalline polymers are transparent. Some examples of noncrystalline, transparent polymers are polycarbonate, acrylics such as poly(methyl methacrylate), and polystyrene. There are some transparent, crystalline polymers, such as polyethylene terephthalate (PET), in which the crystallites are smaller than the wavelength of light. An interesting artifact of polymer crystallinity is the anisotropic nature of the refractive index, which leads to birefringence, as first introduced in Section 6.3.2.1. When the topic of polymer crystallinity was first introduced back in Chapter 1, we presented a photomicrograph of polyethylene spherulites that was produced under cross-polarized light (cf. Figure 1.62). This image is obtained by placing a sample of the polymer between two quarter-wave plate optical filters, or polarizers, which allow the light passing through to vibrate only along one plane. In this way, if the second filter, called the analyzer, is rotated 90◦ from the polarizer, no light will pass through the filter combination, since the plane of vibration allowed through the polarizer is extinguished in the analyzer. However, when a crystalline sample such as polyethylene is placed between the polarizer and analyzer, an interesting pattern is transmitted through the analyzer, which includes some dark regions (no transmission) and light regions (complete transmission). An idealized form of the resulting Maltese cross is shown in Figure 6.103, and an actual image of which was shown in Figure 1.62. This image is the result of birefringence, or refractive index differences in the oriented polymer chains. The index of refraction for most polymers is greater parallel to the chain than normal to the molecular axis. As a result, the refractive index in the tangential direction of the spherulite is greater (see Figure 1.61) than that along the spherulite radius. This birefringence, coupled with the spherical geometry of the spherulite, produces light extinction along the axis of each of the filters, hence the 90◦ angles in the Maltese cross. Birefringence can also be used to analyze polymer samples after melt processing. As we will see in the next chapter, the shear produced in certain molding techniques, such as injection molding, can orient polymer chains in certain parts of the mold, especially near the mold walls, whereas the chains in low-shear regions, such as in the middle of the mold, are not as oriented. Figure 6.104 shows the variation in birefringence, as

670

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Figure 6.103 Schematic of Maltese cross produced by spherulites in crosspolarized filters. Reprinted, by permission, from Strobl, G., The Physics of Polymers, 2nd ed., p. 146. Copyright  1997 Springer-Verlag.

∆n

0.01

0

0 −H/2 +H/2 Position from mold center

Figure 6.104 Variation in birefringence of injection molded polymer of width H . Adapted from P. C. Powell and A. J. Ingen Housz, Engineering with Polymers, p. 314. Copyright  1998 by P. C. Powell and A. J. Ingen Housz.

measured by the refractive index difference along two directions, with position in the mold for an injection-molded polymer. The birefringence is largest at the periphery of the mold, where shear rates are highest and the polymer is most highly oriented. 6.3.3.2 Electroluminescence∗ . In Section 6.3.2.5, we saw that some materials—in particular, semiconductors—can reemit radiation after the absorption of light in a process called photoluminescence. A related type of emission process, which is common in polymer-based semiconductors, called electroluminescence, results when the electronic excitation necessary for emission is brought about by the application of an electric field rather than by incident photons. The electric field injects electrons into the conduction band, and holes into the valence band, which upon recombination emit light.

OPTICAL PROPERTIES OF MATERIALS

671

As we saw in Section 6.1.3.2, most electrically conducting polymers contain polyconjugated structures, such as extended double bonds along the polymer backbone. It is not surprising, then, that electrically conducting polymers such as polyacetylene can also be electroluminescent. The first visible-light-emitting polymer was made from poly(phenylene vinylene), or PPV, the structure for which is shown in Figure 6.105a. PPV exhibits luminescence in the yellow part of the spectrum, and it can be fabricated into a light-emitting diode (LED) by sandwiching a micron-thick layer of PPV between contacts made of indium-tin oxide (a transparent electrode material) on the bottom and cadmium metal on the top. These contact materials are selected to match their electronic bands to that of the polymer. When 12 volts are applied across the contacts, electrons and holes enter the polymer and recombined to generate photons of yellow light. Other polymers with different band gaps can be used to emit light of various wavelengths, such as polythiophene-based compounds (see Figure 6.105b) for the emission of blue light. The applications for polymer LEDs (PLEDs) continue to grow, including paper-thin computer and television displays, which will ultimately require a single polymer that is capable of emitting blue, red, or yellow light, depending upon the applied voltage. The requirements that must be met before PLEDs can become viable candidates for use in flat panel displays include sufficiently high brightness, good photoluminescence profiles, high efficiency (i.e., low operating voltage and current), good color saturation, and long lifetime. A good display device must have at least a brightness of 100 cd m−2 at an operating voltage of between 5 and 15 V and a lifetime of 10,000 h. Moreover, charge conduction in the PLEDs requires electric fields in the range of 2–5 MV cm−1 . 6.3.3.3 Liquid Crystal Displays (LCDs)∗ . Liquid crystalline polymers, first introduced in Section 1.3.6.3, are utilized for a different type of computer and television display, the liquid crystal display (LCD). Most of today’s laptop computers and handheld devices utilize color flat panel displays where the light transmission from the

(a) Poly (p−phenylene vinylene)

S

S S

S

(b) Polythiophere

Figure 6.105 Molecular structure of two electroluminescent polymers. Reprinted, by permission, from P. Ball, Made to Measure, p. 381. Copyright  1997 by Princeton University Press.

672

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

back to the front of the display is modulated by orientational changes in liquid crystal molecules. As illustrated in Figure 6.106, liquid crystalline polymers are sandwiched between two transparent polarizers. The types of liquid crystalline polymers can vary, but generally they are of the cholesteric or twisted nematic type. The LC polymers are initially oriented in the display manufacturing process using a “rubbing process.” The mechanism of alignment is not well understood, however, and is an expensive step in the fabrication process. The transparent substrates between which the liquid crystals are deposited are usually polyimides. Under an applied electric field, the liquid crystal molecules tend to align parallel to each other, resulting in optical anisotropy. The liquid crystal cell is thus a tiny shutter that is controlled by the applied electric field. When the area encompassed on the top, bottom, and sides by electrodes, called a pixel, is turned on, the liquid crystal molecules align, allowing only one plane of the electromagnetic wave (polarized light) to penetrate to the front panel, where it is absorbed in the second polarizer, effectively blocking the light (left image in Figure 6.106). When the field is turned off, the shutter filters the light by rotating its optical plane, as shown on the right in Figure 6.106. Changing bright pixels into colored ones requires yet another filter. As a result, LCDs

Power source

Outgoing light Liquid crystal molecule

Polarizer

Transparent electrode on glass substrate

Incoming light Polarizer

Figure 6.106 Schematic illustration of liquid crystalline polymers sandwiched between polarizers and effect on light in a liquid crystal display. From Jun-ichi Hanna and Isamu Shimizu, Materials in active-matrix liquid-crystal displays, MRS Bulletin, 21(3), 35 (1996). Reproduced by permission of MRS Bulletin.

OPTICAL PROPERTIES OF MATERIALS

673

are very dim, and must be back-illuminated with a lamp, which is an energy-intensive process. LCDs also suffer from a limited viewing angle. These are the primary reasons behind the interest in the electroluminescent materials of the previous section for flat panel displays. 6.3.3.4 Nonlinear Optical Materials∗ . We have seen several examples in the previous sections of how photons can be used to generate and transmit data. Though it is beyond the scope of this text, there are also devices for storing data in an optical fashion. This emerging area of using photons to replace what was previously performed by electrons is called photonics. The integration of photonic and electronic devices is called optoelectronics. In this section, we describe how the optical properties of materials can be used in yet another manner to create devices that can act as switches, multiplexers, and mirrors and even change the wavelength of light. The materials that exhibit these effects and that are used in these photonic and optoelectronic devices are generally referred to as nonlinear optical materials (NLO), because their outputs are not a linear response to some input. We are most familiar with linear responses of materials; for example, the intensity of photons emitted from a tungsten filament in a light bulb increases (more or less) linearly with the power we put into it. But even in this example, nonlinear behavior can occur in the form of saturation, where further increases in power do not result in more output, or in the form of breakdown, when the filament burns out. The diodes described in the preceding sections are nonlinear electronic components—their output currents remain negligibly low until the driving voltages reach a threshold value whereupon the output current increases sharply. The best way to describe a nonlinear optical response is first through some examples, then through some mathematics. Shortly after the invention of lasers, it was observed that a ruby laser beam passing through a quartz crystal produced a faint beam at the laser’s second harmonic—that is, at twice the fundamental frequency of the ruby laser (cf. Section 6.3.2.5). A much simpler-to-perform example of nonlinear optical behavior is found when the polarization of light passing through a crystal is modified upon application of an electric field to the crystal. The origin of both of these NLO effects is related to the change in refractive indices that result by an applied electric field and the modulation of light beams by these field-dependent indices. We have already seen the relationship between refractive index and polarizability, and this is where our mathematical description of NLO effects begins. We can examine how induced polarization behaves as a function of an applied electric field, E , by considering the induced electric dipole moment (cf. Section 6.1.2.2), pe , as a Taylor series expansion in E :

pe = pe,0 + E (∂pe /∂ E )E →0 + 1/2E 2 (∂ 2 pe /∂ E 2 )E →0 + 1/6E 3 (∂ 3 pe /∂ E 3 )E →0 + · · · (6.93) or, more simply, pe = pe,0 + α E + (β/2)E 2 + (γ /6)E 3 + · · ·

(6.94)

where pe,0 is the static dipole of the molecule, α is the linear polarizability [previously called just “polarizability”; cf. Eq. (6.35)], and the higher-order terms β and γ are called the first and second hyperpolarizabilities, respectively. Note that once again, we consider only the magnitude of the electric field, which is actually a vector quantity.

674

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

The terms beyond α E in Eq. (6.94) are not linear in E , so they are referred to as the nonlinear polarization and give rise to NLO effects. You should now see why nonlinear polarization becomes more important at higher electric field strengths, since it scales with higher powers of the field. For most materials, α E > (β/2)E 2 > (γ /6)E 3 , so the first few observations of NLO effects were made prior to the invention of lasers and their associated large electric fields. The bulk polarization, P , which you will recall is the dipole moment per unit volume, can then be expressed as P = P0 + χ (1) E + χ (2) E 2 + χ (3) E 3 + · · ·

(6.95)

where the χ (n) susceptibility coefficients (not to be confused with the magnetic susceptibility of the same symbol) are tensors of order n + 1, and P0 is the intrinsic static dipole moment density of the material. We now consider the polarization induced by an oscillating electric field, such as that found in electromagnetic radiation (light), which can be expressed as E = E0 cos(ωt)

(6.96)

where ω is the frequency of the electric field and t is time. Substitution of Eq. (6.96) into Eq. (6.95) yields P = P0 + χ (1) E0 cos(ωt) + χ (2) E02 cos2 (ωt) + χ (3) E03 cos3 (ωt) + · · ·

(6.97)

Because cos2 (ωt) = 1/2 + 1/2 cos(2 ωt), the first three terms of Eq. (6.97) become P = (P0 + 1/2χ (2) E02 ) + χ (1) E0 cos(ωt) + 1/2χ (2) E02 cos(2 ωt) + · · ·

(6.98)

Equation (6.98) indicates that the polarization consists of a second-order, direct-current (not dependent upon ω) field contribution to the static polarization (the first term), a frequency component, ω, corresponding to the incident light frequency (the second term), and a new, frequency-doubled component, 2 ω (the third term). Thus, if an intense light beam passes through a second-order NLO material, light at twice the input frequency will be produced, as well as a static electric field. The process of frequencydoubling, which explains the first example of NLO behavior from the beginning of this section, is called second harmonic generation (SHG). The oscillating dipole reemits at all of its polarization frequencies, so light is observed at both ω and 2 ω. The process of static field production is called optical rectification. The general condition of second-order NLO effects involves the interaction of two distinct waves of frequencies ω1 and ω2 in an NLO material. In this case, polarization occurs at sum (ω1 + ω2 ) and difference (ω1 − ω2 ) frequencies. This electronic polarization will therefore reemit radiation at these frequencies, with contributions that depend on χ (2) , which is itself frequency-dependent. The combination of frequencies is called sum (or difference) frequency generation (SFG). SHG is a special case of SFG, in which the two frequencies are equal. An applied electric field can also change a material’s linear susceptibility, and thus its refractive index. This effect is known as the linear electro-optic (LEO) or Pockel’s effect, and it can be used to modulate light by changing the voltage applied to a second-order NLO material. The applied voltage anisotropically distorts the electron

OPTICAL PROPERTIES OF MATERIALS

675

density within the material, and the optical beam “sees” a different polarizability and anisotropy of the polarizability. As a result, a beam of light can have its polarization state changed by an amount related to the strength of the orientation of the applied voltage and can travel at a different speed, and possibly in a different direction. The change in refractive index as a function of the applied field is approximated by the general expression 1/Nij2 = 1/n2ij + rij k Ek + sij kl Ek El + · · ·

(6.99)

where Nij are the induced refractive indices, nij are the refractive indices in the absence of the electric field, rij k are the linear or Pockels coefficients, sij kl are the quadratic or Kerr coefficients, and E is the applied field. Subscripts indicate the orientation of the field with respect to a material coordinate system. Equation (6.99) indicates that light traveling through an electro-optic material can be phase- or polarization-modulated by refractive index changes induced by an applied electric field. Devices exploiting this effect include optical switches, modulators, and wavelength filters. An example of an electro-optic switch based on NLO materials is shown in Figure 6.107. The switch is comprised of two parallel waveguides made of NLO materials. The waveguide channels have a different refractive index from the surrounding material. The light can be switched back and forth between the channels by applying and removing a voltage across the bottleneck. In the absence of an electric field, the light traveling through the lower waveguide interacts with the upper waveguide in a nonlinear manner at the bottleneck, causing the light to switch channels. Switching does not occur when an electric field is applied. The electric field polarizes the NLO material and alters the refractive indices of the two channels, such that the nonlinear interaction at the bottleneck is modified and the light stays in the lower waveguide. Many current NLO devices are based upon crystalline materials (such as lithium niobate for the electro-optic switch) and nonlinear optical glasses, but there is intense

Waveguide Light (a)

(b)

Figure 6.107 Schematic illustration of an electro-optic switch. (a) An open circuit causes light to switch channels, and (b) a closed switch keeps light in the same channel. Reprinted, by permission, from P. Ball, Made to Measure, p. 53. Copyright  1997 by Princeton University Press.

676

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Acrylate backbone

F F F F

F

F

F

F

( a)

(a)

NLO side-groups (b)

Figure 6.108 Examples of nonlinear optical polymers (a) fluorinated polyimide and (b) polyacrylate with NLO side groups. Reprinted, by permission, from P. Ball, Made to Measure, p. 55. Copyright  1997 by Princeton University Press.

interest in utilizing polymers for these applications, due to their inherently low densities and potential ease of processing. Some of the polymers that have been used for photonic devices are based upon architectures that have highly polarizable electronic structures. Several polyimides fall into this category (Figure 6.108a), as do polyacrylates and polyurethanes with optically responsive side groups (Figure 6.108b). 6.3.4

Optical Properties of Composites and Biologics∗

Of the three physical properties covered in this chapter, optical properties have the least importance in composite and biological applications. This is not to say that there are no applications of optical properties in composites or biological materials. There are indeed, such as the use of birefringence in the analysis of stress distribution and fiber breakage in fiber–matrix composites [14] and in the development of materials for ophthalmic implants such as intraocular devices [15]. These topics are beyond the scope of this text, however, even as optional information, and introduce no new concepts from a material property standpoint. There are many interesting articles and

REFERENCES

677

books on these topics, however, and the interested reader is encouraged to consult them as these, and many more, new uses for the optical properties of materials are applied to composite and biological materials systems.

REFERENCES Cited References 1. Solymar, L., and D. Walsh, Lectures on the Electrical Properties of Materials, 5th ed., Oxford University Press, Oxford, 1993, p. 425. 2. Sheahan, T., Introduction to High-Temperature Superconductivity, Plenum, New York, 1994, p. 3. 3. Cava, R. J., Oxide superconductors, J. Am. Ceram. Soc., 83(1), 5–28 (2000). 4. Sheahan, T., Introduction to High-Temperature Superconductivity, Plenum Press, New York, 1994, p. 223. 5. Zawodzinski, T. A., C. Derouin, S. Radzinski, R. J. Sherman, V. T. Smith, T. E. Springer, and S. Gottesfeld, J. Electrochem. Soc., 140, 1041 (1993). 6. Schmidt, C. E., V. R. Shastri, J. P. Vacanti, and R. Langer, Stimulation of neurite growth using an electrically-conducting polymer, Proc. Natl. Acad. Sci. USA, 94, 8948–8953 (1997). 7. Sun, A., H. Xu, Z. Chen, L. Cui, and X. Hai, Research on electrical properties of amphiphilic lipid membranes by means of interdigital electrodes, Mater. Sci. Eng. C, 2(3), 159–163 (1995). 8. Hontsu, S., T. Matsumoto, J. Ishii, M. Nakamori, H. Tabata, and T. Kawai, Electrical properties of hydroxyapatite thin films grown by pulsed laser deposition, Thin Sol. Films, 295, 214–217 (1997). 9. Gatteschi, D., P. Carretta, and A. Lascialfari, Molecular magnets and magnetic nanoparticles: new opportunities fo µSR investigations, Physica B, 289–290, 94–105 (2000). 10. Yakhmi, J. V., Magnetism as a functionality at the molecular level, Physica B, 321, 204–212 (2002). 11. Chavan, S. A., J. V. Yakhmi, and I. K. Gopalakrishnan, Molecular ferromagnets—a review, Mater. Sci. Eng. C, 3, 175–179 (1995). 12. Miller, J. S. and A. J. Epstein, Molecular magnets: An emerging area of materials chemistry, in Materials Chemistry: An Emerging Discipline, L. V. Interrante, L. A. Casper, and A. B. Ellis, eds., American Chemical Society, Washington, D.C., 1995, p. 161. 13. Hanna, J., and I. Shimizu, Materials in Active-Matrix Liquid Crystal Displays, MRS Bulletin, March, 35 (1996). 14. Schuster, D. M., and E. Scala, Trans. AIME, 230, 1639 (1964). 15. Obstbaum, S. A., Ophthalmic Implantation, in Biomaterials Science, B. D. Ratner, A. S. Hoffman, F. J. Schoen, and J. E. Lemons, eds., Academic Press, San Diego, 1996, p. 435.

Electrical Properties of Materials Electrical Properties of Bone and Cartilage, C. T. Brighton, J. Black, and S. Pollack, eds., Grune & Stratton, New York, 1979. Solymar, L., and D. Walsh, Lectures on the Electrical Properties of Materials, 5th ed., Oxford University Press, New York, 1993.

678

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

Magnetic Properties of Materials Moorjani, K., and J. M. D. Coey, Magnetic Glasses, Elsevier, Amsterdam, 1984. Westbrook, C., and C. Kaut, MRI in Practice, Blackwell Science, Oxford, 1998.

Optical Properties of Materials Campbell, S. A., The Science and Engineering of Microelectronic Fabrication, Oxford University Press, New York, 1996. Wolf, S., and R. N. Tauber, Silicon Processing for the VLSI Era, Vol. 1, Lattice Press, Sunset Beach, CA, 1986. Fox, M., Optical Properties of Solids, Oxford University Press, New York, 2001.

PROBLEMS Level I

6.I.1 What will be the resistance of a copper wire 0.08 in. in diameter and 100 ft long if its resistivity is 1.7 µ · cm? 6.I.2 A maximum resistance of 1  is permitted in a copper wire 25 ft long. What is the smallest wire diameter that can be used? 6.I.3 What is the electrical conductivity of iron (a) at room temperature? (b) at 212◦ F? 6.I.4 Silicon has a density of 2.40 g/cm3 . (a) What is the concentration of the silicon atoms per cubic centimeter? (b) Phosphorus is added to the silicon to make it an n-type semiconductor with a conductivity of 1 mho/cm and an electron mobility of 1700 cm2 /V-s. What is the concentration of the conduction electrons per cubic centimeter? 6.I.5 (a) How many silicon atoms are there for each conduction electron in problem ˚ and there are eight atoms 6.I.4? (b) The lattice constant for silicon is 5.42 A, per unit cell. What is the volume associated with each conduction electron? 6.I.6 Germanium used for transistors has a resistivity of 2  · cm and an electron “hole” concentration of 1.9 × 1015 holes/cm3 . (a) What is the mobility of the electron holes in the germanium? (b) What impurity element could be added to germanium to create electron holes? 6.I.7 Calculate the mobility of electrons in Cu. The resistivity of Cu is 1.72 × 10−8  · m at 25◦ C and its density is 8.9 g/cm3 . Assume each copper atom donates one valence electron to the conduction band. 6.I.8 A coil of wire 0.1 m long and having 15 turns carries a current of 1.0 A. (a) Compute the magnetic induction if the coil is within a vacuum. (b) A bar of molybdenum is now placed in the coil, and the current adjusted to maintain the same magnetic induction as in part (a). Calculate the magnetization. 6.I.9 Why is there a maximum power loss when the dielectric is changing most rapidly as a function of temperature and frequency?

PROBLEMS

679

6.I.10 Crown glass has a refractive index of 1.51 in the visible spectral region. Calculate the reflectivity of the air–glass interface, and determine the transmission of a typical glass window. 6.I.11 Look up the refractive indices for fused silica and dense flint glass, and calculate the ratio of their reflectivities. Cite the source of your information. Level II

6.II.1 At room temperature, the temperature dependence of electron and hole mobilities for intrinsic germanium is found to be proportional to T −3/2 for temperature in degrees Kelvin. Thus, a more appropriate form of Eq. (6.31) is ′′

σ =C T

−3/2



−Eg exp 2kB T



where C ′′ is a temperature-independent constant. (a) Calculate the intrinsic electrical conductivity using both Eq. (6.31) and the equation above for intrinsic germanium at 150◦ C, if the room temperature electrical conductivity is 2.2 ( · m)−1 . (b) Compute the number of free electrons and holes for intrinsic germanium at 150◦ C assuming this T −3/2 dependence of electron and hole mobilities. 6.II.2 Assume there exists some hypothetical metal that exhibits ferromagnetic behavior and that has a simple cubic structure, an atomic radius of 0.153 nm, and a saturation flux density of 0.76 tesla. Determine the number of Bohr magnetons per atom for this material. 6.II.3 The formula for yttrium iron garnet (Y3 Fe5 O12 ) may be written in the form Y3 c Fe2 a Fe3 d O12 , where the superscripts a, c, and d represent different sites on which the Y3+ and Fe3+ ions are located. The spin magnetic moments for the Y3+ and Fe3+ ions positioned in the a and c sites are oriented parallel to one another and antiparallel to the Fe3+ ions in d sites. Compute the number of Bohr magnetons associated with each Y3+ ion, given the following information: (1) Each unit cell consists of eight formula (Y3 Fe5 O12 ) units; (2) the unit cell is cubic with an edge length of 1.2376 nm; (3) the saturation magnetization for this material is 1.0 × 104 A/m, and (4) assume that there are 5 Bohr magnetons associated with each Fe3+ ion. 6.II.4 The detectors used in optical fiber networks operating at 850 nm are usually made of silicon, which has an absorption coefficient of 1.3 × 105 m−1 at this wavelength. The detectors have coating on the front face that make the reflectivity at the design wavelength negligibly small. Calculate the thickness of the active region of a photodiode designed to absorb 90% of the light. Level III

6.III.1 The magnetocaloric effect (Scientific American, May 1998, p. 44) relies on the ability of a ferromagnetic material to heat up in the presence of a magnetic field and then cool down once the field is removed. When a ferromagnet is

680

ELECTRICAL, MAGNETIC, AND OPTICAL PROPERTIES OF MATERIALS

placed in a magnetic field, the magnetic moments of its atoms become aligned, making the material more ordered. But the amount of entropy must be conserved, so the atoms vibrate more rapidly, raising the material’s temperature. When the material is taken out of the field, the material cools. Water, or some other heat transfer fluid, can be cooled down by running it through the ferromagnetic material as it cools. (a) The magnitude of the magnetocaloric effect reaches a maximum at the Curie temperature of the ferromagnet. Based upon the above description of the application and the Curie temperature, which of the following three materials will make the best ferromagnetic material for a typical household refrigeration unit? Justify your selection.

Material

Saturation Magnetizationa at Room Temperature

Curie Temperature (K)

Melting Point (K)

Iron Gadolinium Nickel

1707 1090 485

1043 289 631

1811 1585 1728

a

Maximum possible magnetization that results when all magnetic dipoles are aligned with the external field.

(b) The amount of refrigeration is also determined by the strength of the applied magnetic field. Superconducting magnets are used to generate massive magnetic fields that cause large magnetizations in the ferromagnetic inductor. Below are three candidate materials for the superconducting magnet. Based on what you know about superconductors, which of these three materials will make the best superconducting magnet for this application? Justify your selection. Material Bi2 Sr2 Ca2 Cu3 O10 Gadolinium NbTi

6.III.2

Tc , K

Hc , tesla

110 5.9 9.5

100 0.0001 10

A team of chemical engineers [Proc. Natl. Acad. Sci. USA, 94, 8984 (1997)] has shown that oxidized polypyrrole supports the growth of structural support cells needed for nerve regeneration in rats. In particular, the team showed that rat nerve cells responded to electrical stimulus from the polymer by producing extensions called neurites that are twice as long as neurites produced in the absence of the stimulus. (a) Calculate the current generated in a 10-mm-long, 10.0 × 10−6 -m-diameter polypyrrole filament that is subjected to a electric field gradient of 12 V. Assume that the conductivity of oxidized polypyrrole is 1.5 × 107 −1 m−1 . (b) One drawback for this application is that polypyrrole is very fragile and not biodegradable. Suggest a method for evaluating the long-term biocompatibility of polypyrrole fibers in vivo.

CHAPTER 7

Processing of Materials

7.0

INTRODUCTION

With an understanding of the structure and properties of engineering materials now firmly in place, we can discuss how these materials can be formed or fabricated into useful products and components. Most of the important processing methods are described here, with little or no distinction made between microscale and macroscale processes—for example, processes that form both integrated circuits and components for highway bridges are described here. The common thread is that all the chemical and physical phenomena needed to introduce these processing techniques have already been described in the previous chapters. By the end of this chapter, you should be able to: ž ž ž ž ž ž ž ž ž ž ž

7.1

Identify and describe different types of metal processing techniques. Calculate the force imparted on a workpiece in forging operations. Calculate the power and penetration depth of metals heated by induction heating. Identify a proper milling technique for the preparation of metallic and ceramic powders of a desired size. Calculate pressures and identify the neutral point in rolling operations. Identify and describe different types of ceramic slurry processing techniques. Differentiate between kinetic and transport rate limitation in the vapor phase processing of ceramics. Identify and describe different types of continuous and batch polymer processing techniques. Use equations of change to derive simple design equations for flow and pressure drop in polymer processing techniques. Differentiate between pultrusion, resin transfer molding, and filament winding methods for fiber-reinforced composite production. Describe the benefits of plasma processing to improvement of biocompatibility in biologics. PROCESSING OF METALS AND ALLOYS

Despite the relatively recent development of polymers and composites as structural materials, metals continue to be the dominant group of engineering materials for many An Introduction to Materials Engineering and Science: For Chemical and Materials Engineers, by Brian S. Mitchell ISBN 0-471-43623-2 Copyright  2004 John Wiley & Sons, Inc.

681

682

PROCESSING OF MATERIALS

applications, including the automotive, aerospace, and construction industries. This is due, in part, to their ease in component fabrication, or formability. In this section, we describe three of the most common and widely utilized metal-processing techniques: casting, wrought processing, and powder metallurgy. The topic of metal joining (e.g., welding), albeit an important one, is beyond the scope of this text. 7.1.1

Casting

In the context of metals processing, casting is the process of melting a metal and solidifying it in a cavity, called a mold, to produce an object whose shape is determined by mold configuration. Almost all metals are cast during some stage of the fabrication process. They can be cast either (a) directly into the shape of the component or (b) as ingots that can be subsequently shaped into a desired form using processes that are described later in this section. Casting offers several advantages over other methods of metal forming: It is adaptable to intricate shapes, to extremely large pieces, and to mass production; and it can provide parts with uniform physical and mechanical properties throughout. From an economic standpoint, it would be desirable to form most metal components directly from casting, since subsequent operations such as forming, extruding, annealing and joining add additional expense. However, since the growth rate of crystalline phases is much higher in the liquid state than in the solid state (cf. Sections 2.1.5 and 3.2.1), the microstructure of as-cast materials is much coarser than that of heat-treated or annealed metals. Consequently, as-cast parts are usually weaker than wrought or rolled parts. Casting is generally only used as the primary fabrication process when (1) the structural components will not be subjected to high stresses during service; (2) high strength can be obtained by subsequent heat treatment; (3) the component is too large to be easily worked into shape; or (4) the shape of the component is too complicated to be easily worked into shape. Let us more closely examine the important steps in the casting process and see how some properties can be related to the important processing parameters in each step. 7.1.1.1 Melting. The initial step in the casting operation is melting. Melting can be achieved through any number of well-established techniques, including resistance heating elements or open flame, but there are several alternative methods that offer distinct advantages in the melting of certain metals. For example, metals such as zirconium, titanium, and molybdenum are melted in water-cooled copper crucibles by arc melting, in which a direct or alternating current is applied across an electrode and the material in the crucible, called the charge. The electrode may be nonconsumable when made of carbon or tungsten, or consumable when made of the metal to be melted. An arc is formed between the electrode and charge, and the energy imparted to the charge causes it to melt. This process is carried out under a vacuum, or in an inert atmosphere such as helium or argon, to prevent contamination with oxygen or nitrogen. In a process known as induction melting, the metal is placed in a container within an electrical coil that produces a high-frequency alternating magnetic field (see Figure 7.1). The alternating current flowing through the induction coil generates an alternating magnetic field that produces a current in the metal, called a workpiece. Heat is generated within the metal due to the passage of these secondary currents induced in the coils by electromagnetic induction (cf. Section 6.2.1.1). The electrical energy through the coil is thus converted to magnetic energy, which is in turn converted back to heat in the

PROCESSING OF METALS AND ALLOYS

Steel shell

Insulating refractory

683

Crucible lid

Crucible

Base block Water-cooled primary induction coil (embedded in refractory cement)

Figure 7.1 Schematic diagram of induction furnace and crucible used in the melting of copper. Reprinted, by permission, from Metals Handbook, pp. 23–60. Copyright  1985 by the American Society for Metals.

workpiece due to the induced secondary electrical currents called eddy currents. The current density in the workpiece is dictated by the skin effect, in which most of the eddy currents occur on the workpiece surface. Toward the center the current density decreases exponentially. In the center of the workpiece the current flow is nearly zero. The heat develops in the workpiece itself; there is no need for a transmission medium, such as air, water, or a mechanical connection. As the outer portions of the metal melt, mixing begins and thermal conduction eventually results in the melting of the entire workpiece. The power imparted to the workpiece or melt, P , is a function of the magnetic susceptibility of the metal, χ, and the frequency of the applied field, ω:  P = kI 2 (1 + χ)ρω (7.1) where P is in watts, k is a constant, I is the current in the induction coil (in amps), ρ is the specific resistivity of the metal in  · mm2 / m, and ω is in Hz. As mentioned above, the induced currents occur at the outer of portions of the metal, as given by the penetration depth, δ, of the currents (in mm)  ρ δ = 503 (7.2) (1 + χ)ω Obviously, the melt or workpiece must be larger in diameter than the penetration depth. The induction process, too, is often carried out under vacuum or inert gas atmosphere.

684

PROCESSING OF MATERIALS

The advantage of induction melting is the ability to melt cleanly and quickly. It is applicable to virtually any metal or alloy, even if they are not inductively suscepting, since it is possible to use the crucible as a susceptor as well. Furthermore, once the metal is molten, the induction lines from the secondary currents stir the melted charge and improve melt homogeneity. Other techniques such as electron beam melting and plasma flame melting are used for highly refractory metals, such as tantalum. These too are often performed in vacuum. The use of vacuum in melting operations is used not only to prevent reactions such as oxidation, but to prevent, and even remove, dissolved gases in the metals. Most metals dissolve considerable amounts of gases in their liquid state. The solubility of these gases is much less in the solid state than in the liquid state, and as the metal solidifies, gases can escape from the melt, forming blowholes and pinholes in the metal ingot. This results in porosity than can adversely affect properties of the metal or alloy. Oxidation is not always to be avoided in the melting process. It is an important step in the making of steel, in which pig iron, the raw product of iron production, is re-melted in a blast of oxygen. The pig iron contains about 4 wt% carbon, which is converted to carbon monoxide in the presence of the oxygen. The furnace used to melt iron and form steel is thus termed a blast furnace, due to the blast of oxygen that is provided. The amount of oxygen is carefully controlled, so that not all of the carbon is removed, but just enough to form steel of the desired carbon content (cf. Figure 2.8 and Table 2.4). The oxygen also reacts to form oxides of highly reactive trace impurities in the pig iron, such as silicon, aluminum, and manganese. These oxides float to the top of the melt, since they are less dense than the molten metal, and form a product called slag that can be skimmed from the top of the melt. An additional benefit of the oxygen blast is that these reactions with carbon and trace impurities produce an enormous amount of energy that can be used to melt the iron. Many modifications have

HISTORICAL HIGHLIGHT

The physical limitations of cast-iron cannons in the Crimean War (1854–1856)—in which England, France, Turkey, and Sardinia defeated Russia for the domination of southeastern Europe—were a turning point in the history of steel. Those limitations caught the attention of Henry Bessemer, an Englishman who had already enjoyed considerable commercial success for inventing the “lead” of “lead pencils.” Bessemer’s breakthrough in 1855 was to drop the standard approach, which involved making steel by heating carbon-free wrought iron with carbonbearing fuel. He tried to turn the process around by starting with carbon-rich pig iron and using oxygen in an air blast to get the excess carbon out. Iron and steel became the stuff of big and visible things—ships, bridges, buildings,

even the Eiffel Tower. These metals, particularly steel, were the harbingers of skyscrapers and audacious urban skylines. In 1874 James Eads used steel to build a giant arched bridge across the Mississippi River in St. Louis. Some 50,000 tons of steel went into the building of a railway bridge over the Firth River in Firth, Scotland. For the enormous suspension cables that would hold up his Brooklyn Bridge, John Roebling in 1883 could choose steel rather than iron. In 1890 the second Rand-McNally Building in Chicago was the first building to be framed entirely in steel. Hundreds of ever higher buildings and skyscrapers followed. Steel also was showing up in hundreds of smaller everyday items like food cans, home appliances, automobiles, and road signs. Adapted from Stuff, I. Amato, pp.42–46

PROCESSING OF METALS AND ALLOYS

685

since been made to this process, and may variations exist, including the formation of pig iron from iron ore (Fe2 O3 ). The chemistry of the steel-making operation is quite complex, and the interested reader is referred to the metallurgy texts cited at the end of this chapter for further information. 7.1.1.2 Forming Operations. In principle, the molten metal can then be poured into a mold of the desired shape and become solidified to form a component. As mentioned above, and as described further in the next section, solidification from casting operations usually leads to undesirable crystal structures, such that the most common operation after melting is the formation of a large metal “blank” called an ingot. Thus, there are two main categories of metal casting processes: ingot casting and casting to shape. Ingot casting makes up the majority of all metal castings and can be separated into three categories: static casting, semicontinuous or direct-chill casting, and continuous casting. Static ingot casting simply involves pouring molten metal into a permanent, refractory mold with a removable bottom or plug in the bottom, as schematically illustrated in Figure 7.2. Ingots have a slight taper toward the bottom of the mold that facilitates their removal and aids in the solidification process. The plug in the bottom of the mold allows the metal to be drained from the mold upon remelting, if desired. Semicontinuous ingot casting is primarily employed in the production of cast aluminum. In this process, the molten metal is transferred to a water-cooled permanent mold that has a movable base mounted on a long piston (see Figure 7.3a). As the metal solidifies, the piston is moved downward and more metal continues to fill the reservoir (see Figure 7.3b). The piston is allowed to move its entire length, at which point the process is stopped. Continuous casting is used in the steel and copper industries. In this process, molten metal is delivered to a permanent mold in much the same way as semicontinuous

Hot top

Metal Refractory

Plug

Figure 7.2 Schematic diagram of an ingot in a refractory mold. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

686

PROCESSING OF MATERIALS

liquid metal

base

water-cooled mold

liquid metal

water spray

piston

base

piston

(a)

(b)

Figure 7.3 Schematic illustration of semicontinuous casting process in which (a) molten metal solidifies in a water-cooled mold with a movable base and (b) the base is moved downward so more metal can be poured into the reservoir. Reprinted, by permission, from the McGraw-Hill Encyclopedia of Science and Technology, Vol. 11, 8th ed., p. 31. Copyright  1997 by McGraw-Hill.

casting. However, instead of the process ceasing after a certain length of time, the solidified ingot is continually cut into lengths and removed from the process as quickly as it is cast. This method has many economic advantages over the more conventional ingot casting techniques. Casting to shape is generally classified according to the molding process, molding material, or method of feeding melt to the mold. There are four basic types of casting to shape processes: sand casting, permanent mold casting, die casting, and centrifugal casting. In sand casting, the mold consists of a mixture of sand grains, water, clay, and other materials, which are compacted around a pattern to 20–80% of their mixture bulk density. The two halves of the mold, called the cope and drag, are clamped together to form internal cavities, into which the metal is cast. When the molds are made of cast iron, steel, or bronze, the process is called permanent mold casting because the molds can be re-used. In this case, the mold cavity is formed by machining. This process is used to cast iron and nonferrous alloys, and it has advantages over sand casting such as smoother surface finish, closer tolerances, and higher production rates. A further development of the permanent molding process is die casting. Molten metal is forced through a plate with an opening of a desired geometry known as a die, under pressures of 0.7–700 MPa (100–100,000 psi). Two basic types of die-casting machines are hot-chamber and cold-chamber. In the hot-chamber machine, a portion of the molten metal is forced into a cavity at pressures up to 14 MPa (2000 psi). The process is used for casting low-melting-point alloys such as lead, zinc, and tin. In the cold-chamber process (see Figure 7.4), the molten metal is ladled into the injection cylinder and forced into the cavity under pressures that are about 10 times those in the hot-chamber process. High-melting-point alloys such as aluminum-, magnesium-, and copper-based alloys are used in this process. Die casting has the advantages of high production rates, high quality and strength, excellent surface finish, and close

PROCESSING OF METALS AND ALLOYS

687

split die molten metal poured in here piston

chamber

Figure 7.4 Schematic illustration of cold chamber die casting process. McGraw-Hill Encyclopedia of Science and Technology, Vol. 11, 8th ed., p. 33. Copyright  1997 by McGraw-Hill.

tolerances. Finally, inertial forces of rotation distribute molten metal into the mold cavities during centrifugal casting. The rotational speed in centrifugal casting is selected to give 40 to 60 gravitational constants of acceleration. Dies may be made of forged steel or cast iron. Other types of casting processes include evaporative casting and zero gravity casting. 7.1.1.3 Solidification. When the ingot or casting solidifies, there are three main possible microstructures that form (see Figure 7.5). We will describe here only the final structures; the thermodynamics of the liquid–solid phase transformation have been described previously in Chapter 2. The outside layer of the ingot is called the chill zone and consists of a thin layer of equiaxed crystals with random orientation.

Chill zone

Columnar zone

Equiaxed Zone

Columnar zone

Chill zone

Figure 7.5 Schematic illustration of crystallization zones in a solidified metal ingot. From The Nature and Properties of Engineering Materials, 2nd ed., by Z. Jastrzebski. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

688

PROCESSING OF MATERIALS

This is due to the high degree of subcooling of the melt adjacent to the mold wall. Recall from Eq. (3.43) that a high degree of undercooling creates a driving force for a high nucleation rate. Many nuclei are formed at the mold wall, and a fine grain size results. The second zone, which constitutes the majority of the ingot, consists of elongated grains, often of dendritic shape, having great length and oriented parallel to the heat-flow direction in the ingot. This is called the columnar zone. Finally, the central zone of the ingot is the last to solidify, resulting in a much lower heat transfer rate and a coarser, equiaxed grain size. The relative amounts of grains in each zone can be controlled by the rate of solidification (cooling) and the composition of the alloy. The cooling rate is determined by the thermal properties of the mold material, the size of the ingot, and the temperature of the melt. If the melt temperature is too high, the chill zone may completely disappear due to redissolution of the chilled grains. In pure metals, as well as in alloys with very little supercooling, the columnar structures may progress to the center of the ingot, where the grains advancing from the opposite wall meet. Thus, no equiaxed zone occurs. At very high cooling rates, the chill zone may extend throughout the ingot, producing a fine-grained, equiaxed microstructure. This usually occurs for small ingots and in cold metal molds. The formation of the central equiaxed zone is caused by a process called constitutional supercooling. This enhances nucleation in the melt at the center of the casting prior to the advance of the solid interface. The growth of the equiaxed zone can also be promoted by adding some nucleating agents. In general, columnar growth dominates in pure metals, while equiaxed crystallization prevails in solid–solution alloys. With increasing amounts of solute atoms, and with increasing freezing range of the solid alloy, the equiaxed crystal zone increases, and in some cases the columnar growth can be completely eliminated. In the extreme of cooling rates, nucleation and crystallization can be completely eliminated, leading to amorphous metals and alloys. This is the topic of the next section. Some common phenomena that occur during solidification are segregation and shrinkage. During the solidification process, variations in the concentration can occur in certain regions of alloys that result in segregation. Two main types of segregation can occur: macrosegregation and microsegregation. In macrosegregation, compositional changes occur over relatively large dimensions as a result of gravity, normal, or inverse segregation. Gravity segregation is caused by density differences in melt components. In normal segregation, the solute is rejected at an advancing solid–liquid interface because of different diffusion rates in solid and in liquid. Inverse segregation is the result of outward movement of the impurity-enriched interdendritic liquid, thereby increasing the average impurity content at the head of the ingot. Microsegregation refers to compositional variations over relatively small distances. Insoluble foreign particles can be trapped during solidification, resulting in segregation at the grain boundaries. Just as polymers undergo volume changes upon cooling (cf. Figure 1.67), so, too, do metals. The resultant shrinkage is equal to the difference in volume between the liquid at the casting temperature and the solid metal at room temperature. The total change in volume is the combined effect of the contraction of the liquid when cooled from its casting temperature to its freezing point, the contraction during freezing from liquid to solid, and the contraction of the solid metal when cooled from its freezing point to room temperature (cf. thermal expansion in Section 5.1.3.1). Shrinkage may cause the formation of cavities in the ingot and adversely affect the properties of the

PROCESSING OF METALS AND ALLOYS

689

Pipe

Centerline shrinkage

Figure 7.6 Schematic illustration of piping in a solidified metal ingot. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

metal. Such cavities, particularly in the extreme case of piping (see Figure 7.6), can cause considerable damage to the ingot since air can penetrate into the cavity and cause oxidation. The presence of oxide layers inside the ingot will prevent proper welding of the cavities during subsequent rolling or forging operations. In order to minimize piping, molds with tapered walls as shown in Figure 7.2 are used, which permit the liquid metal to penetrate into any cavities or pores that may have already formed in the solidified parts of the ingot. Shrinkage can also be minimized in cast parts through use of a riser, which is a volume added to the casting to provide metal as shrinkage occurs. The relationship between the processing history and the solidification event is illustrated in Figure 7.7 for a cast component. Solidification begins with the formation of dendrites on the mold walls, which grow with time into the liquid and eventually consume the entire liquid volume. In alloys, it is the growth of dendrites that leads to segregation. In addition, the dendrites provide opportunities for gases to become trapped and form porosity. Most or all of these defects can be eliminated by processing in a microgravity environment. Macrosegregation, for example, is driven by gravity effects in which solute-rich materials sink in response to density differences. 7.1.1.4 Rapid Solidification. In Section 6.2.1.3, we saw the relationship between grain size, magnetic domain size, and magnetic properties in ferrous materials. In the previous section, we saw the effect of cooling rate on grain size in metals. You would expect, then, that there would be a strong relationship between solidification processes and magnetic properties in ferrous alloys, and this is indeed the case. We also saw that increased solidification rates at the walls of a mold can produce a finegrained crystalline structure due to increased nucleation rates. But in the extreme, even nucleation rates drop off at sufficient subcooling (cf. Figure 3.15), such that it is possible, in theory, to produce amorphous metals and alloys, also known as metallic glasses, with high enough cooling rates. Such amorphous materials are formed by a class of techniques known collectively as rapid solidification processing, in which solidification rates, or quench rates as they are called for these processes, are on the

690

PROCESSING OF MATERIALS

riser

core

air liquid metal

liquid metal

0

mold preparation complete

10

begin pouring

20 30

end pouring

40

solidification begins (42 s)

50 60 70

90 liquid 100 solid dendrite

110

time, s

80

1. dendrites form and grow 2. gas bubbles form and attempt to float free 3. shrinkage occurs as liquid → solid 4. alloying elements segregate between dendrite arms

120 gas bubble

shrinkage cavity

130 140

solidification ends

150 second-phase particles

160 170

Figure 7.7 Solidification history for a cast-to-shape component. McGraw-Hill Encyclopedia of Science and Technology, Vol. 11, 8th ed., p. 35. Copyright  1997 by McGraw-Hill.

order of 105 to 106 ◦ C/s. It is not yet possible to produce amorphous pure metals, since the minimum cooling rates through the liquid–solid transition must be at least 1014 ◦ C/s, so metallic glasses always contain at least two or more different kinds of atoms. Although rapid solidification may not produce a truly amorphous (noncrystalline) material for some alloy compositions, crystallite sizes of rapidly solidified crystalline

PROCESSING OF METALS AND ALLOYS

Table 7.1 Alloys

691

Composition of Some Amorphous Ferrous

Composition (wt%, balance Fe) B

Si

Cr

Ni

Mn

P

20 10 28 6

— 10 — —

— — 6 —

— — — 40

— — 6 —

— — — 14

alloys are characteristically in the range of 500 nm or below, compared with 50 µm for traditionally processed alloys. The absence of grain boundaries in alloys such as iron–silicon improve their magnetization properties and help make them especially attractive for soft-magnet applications such as transformer cores. Boron is another common alloying element in ferrous alloys that assists in the formation of noncrystalline structures. The compositions of some common amorphous ferrous alloys produced by rapid solidification techniques are listed in Table 7.1. Other known glass-forming metallic systems include alloys of transition metals or noble metals that contain about 10–30% semimetals, such as the iron–boron systems above, platinum–phosphorus, and niobium–rhodium, alloys containing group II metals such as magnesium–zinc, and alloys of rare-earth metals and transition metals such as yttrium–iron and gadolinium–cobalt. With further solute substitution, the stability and glass-forming ability can be drastically enhanced. Ternary alloy glasses such as palladium–copper–silicon and platinum–nickel–phosphorus have been prepared as cylindrical rods of 2.5 mm in diameter at quench rates of 102 ◦ C/s or less. The X-ray diffraction pattern (cf. Section 1.1.2) of an amorphous alloy is very similar to that of the same alloy in the liquid state. From the XRD pattern, it can be determined that the coordination number in metallic glasses is close to 12, just as in the close-packed crystalline structure. The high quench rates necessary to achieve the noncrystalline structures are primarily achieved through contacting the molten alloys with a surface that has a high thermal conductivity and is maintained at a low temperature. In a process called splat-quenching, the molten metal is propelled by means of a shock tube onto copper substrates with high thermal conductivity. Ribbons, or tapes, of metallic glasses can be formed by feeding a continuous jet of liquid alloy on the outside rim of a rapidly rotating cylinder. The molten alloy solidifies rapidly into a thin ribbon, about 50 µm thick, that is ejected tangentially to the rotating cylinder at rates as high as 2000 m/min. The width of the ribbon is limited to a few millimeters. Ribbons up to 15 cm in width can be produced by using a planar flow casting method, as illustrated in Figure 7.8. By injecting a stream of melt through a nozzle into a rotating drum containing water instead of onto a rotating cylinder, wires with circular cross sections can be formed. The centrifugal force ensures the formation of a water layer on the inner surface of the drum. The molten jet solidifies rapidly into a wire that is collected at the bottom of the drum. The wire has a uniform cross section with diameter ranging from 50 to 200 µm. The metallic glass wires show excellent mechanical and magnetic properties, comparable to or surpassing those of corresponding ribbons.

692

PROCESSING OF MATERIALS

nozzle

slotted nozzle

metallic glass strip

melt

substrate substrate (a)

(b)

Figure 7.8 Schematic illustration of (a) nozzle-substrate orientation and (b) cross section of the nozzle substrate in planar flow casting. McGraw-Hill Encyclopedia of Science and Technology, Vol. 11, 8th ed., p. 57. Copyright  1997 by McGraw-Hill.

7.1.2

Wrought Metals and Alloys

The ingots produced by casting can either be used as cast, or further processed into usable geometries such as bars, rods and sheets. Those metals and alloys that are manufactured by plastic deformation in the solid state are referred to as wrought metals and alloys.∗ The mechanical working of metals during wrought processing improves the properties of the wrought metals over that of the cast products. Plastic deformation, when carried out below the recrystallization temperature (cf. Section 5.1.2.3), results in work hardening, which improves the mechanical properties of the metal. In all wrought processes, the flow of metal is caused by application of an external force or pressure that pushes or pulls a piece of metal or alloy through a metal die. The pressure required to produce plastic flow is determined primarily by the yield stress of the material (cf. Section 5.1.4.3) which, in turn, controls the load capacity of the machinery required to accomplish this desired change in shape. The pressure, P , used to overcome the yield stress and cause plastic deformation is given by P = σy ε

(7.3)

where σy is the yield stress of the material and ε is the strain during the deformation. It is often useful to utilize the true strain in this relationship [see Eq. (5.31)] rather than the engineering strain. The pressure calculated in this way is generally much lower than that required in the actual process due to losses caused by friction, work hardening, and inhomogeneous plastic deformation. For example, the pressure predicted in Eq. (7.3) is only 30–55% of the working pressure required for extrusion, 50–70% of that required for drawing, and 80–90% of the pressure required for rolling. The actual work requirement is the sum of (a) the work due to plastic deformation and (b) the losses mentioned above. The mechanical working of metals can be accomplished by various processes, but the methods of forging, rolling, and extrusion will be described here. ∗

There is no participial form of the word “wrought” in common use that is analogous to the term “casting,” so the modified gerund wrought processing, albeit awkward, is generally used.

PROCESSING OF METALS AND ALLOYS

693

7.1.2.1 Forging. Forging is the process of applying large forces, mainly compressive forces, to cause plastic deformation to occur in the metal. The use of compressive forces means that tensile necking and fracture are generally avoided. Forging can be used to produce objects of irregular shape. There are two main categories of forging: pressing and hammering. Pressing is the relatively slow deformation of metal to the required shape between mating dies. This method is preferred for homogenizing large cast ingots because the deformation zone extends throughout the cross section. All presses are composed of some basic components. These include a ram for applying the compressive force, a drive to move the ram up and down, power to the drive, a bed to rest the workpiece upon, and a frame to hold the components. Most press forging is done on upright hydraulic presses or mechanical presses. There are a variety of frame geometries and types of ram drives that vary depending upon the pressing application. A press is rated on the basis of the force (typically in tons) it can deliver near the bottom of a stroke. Pressures (in tons/in.2 of projected area) have been found to be 5–20 for brass, 19–20 for aluminum, 15–30 for steel, and 20–40 for titanium. The force the press must deliver is equal to the unit pressure times the projected area. Presses in excess of 50,000 tons have been built for forging large and complex parts. Hammering involves more rapid deformation than pressing and is mainly restricted to the surface region of the metal. Large forgings require very large forces and usually involve elevated temperatures. The most common forging hammers are steam or air operated. An example of a steam forging hammer is shown in Figure 7.9. A related operation is drop forging, in which a drop hammer forms parts with impression or cavity dies. Stock metal in the form of bars, slugs, or billets is placed in a cavity in the bottom half of a forging die on the anvil of a drop hammer. The upper half of the die is attached to the hammer, which falls on the stock, shaping it into the desired geometry. Generally, a finished forging cannot be formed in one blow, so most dies have several impressions, and the workpiece is transferred from one to the next between blows until the finished product is attained. 7.1.2.2 Rolling. In the process of rolling, the metal is continuously drawn between rotating rollers by the friction forces between the surfaces of the rolls and the metal, as illustrated in Figure 7.10, where hf < h0 . The process can be run either hot or cold, and it is much more economical than forging since it is faster, consumes less power, and produces items of a uniform cross section in a continuous fashion. In cold rolling, it is possible to attain production speeds of over 1500 m/min for thin strips of metal. Temperatures in the hot rolling process are similar to those in forging, namely 400–450◦ C for aluminum alloys, 820◦ C for copper alloys, 930–1260◦ C for alloys steels, 760–980◦ C for titanium alloys, and 980–1650◦ C for refractory metals. A number of roll arrangements are possible, as illustrated in Figure 7.11. Unlike other wrought processes, rolling can be used without a die. Deformations are restricted to a small volume at any given time, so the loads are relatively low. The metal is subjected to a longitudinal tension force that, combined with the normal force, produces a shear stress, τ , which causes deformation

τ =f ×P

(7.4)

where P is the normal pressure applied by the roller and f is the coefficient of friction between the metal surface and the roller surface. In hot rolling, f may be as much as

694

PROCESSING OF MATERIALS

Safety head

Operating valve Steam cylinder Throttle valve Exhaust Throttle control

Piston rod

Frame Guide Motion

Ram Upper die Lower die Anvil cap

Anvil

Figure 7.9 Schematic illustration of a single frame steam forging hammer. Reprinted, by permission, from Doyle, L.E., Manufacturing Processes and Materials for Engineers, 2nd ed., p. 254. Copyright  1969 by Prentice-Hall, Inc.

0.7, whereas in cold rolling it generally ranges from 0.02 to 0.3. The normal pressures entering, Pe , and exiting, Px , the neutral point can be determined at any point of contact along the arc formed between the roller and the sheet by Pe = σ ′

h exp[f (H0 − H )] h0

(7.5)

h exp(f H ) hf

(7.6)

and Px = σ ′

where f is again the coefficient of friction, σ ′ is the plane strain flow stress at a particular point in the roll gap where the instantaneous thickness is h, h0 is the entering thickness, hf is the exiting thickness, and H is a parameter given by     R R (7.7) tan−1 θ H =2 hf hf

PROCESSING OF METALS AND ALLOYS

695

Example Problem 7.1

The ability of a hammer to deform metal depends on the energy it is able to deliver on impact. Consider a steam hammer (see Figure 7.9) that has a falling weight of 200 lb and a steam bore, d, equal to 12 in. Assume that the mean effective steam pressure, P, is 80 psi and that the stroke is 30 in. If the hammer travels 1/8 in. into the metal after striking it, determine the average force exerted on the workpiece. You should be able to do this without an equation, through the application of some dimensional analysis.   πd2 × P = 9050 lbf Answer : Steam force = 4 Total downward force = Steam force + Hammer weight = 9050 + 2000 = 11,050 lbf Energy in a blow = Force × Distance = 11,050 lbf × 30 in. = 331,500 in.-lbf Average force exerted = 331, 500/0.125 = 2,652,000 lbf (1326 tons)

R q

V0 N h0

Vf hf

N

Figure 7.10 Schematic diagram of roll gap region in the rolling of metals.

696

PROCESSING OF MATERIALS

(a)

(b)

(c )

(d )

(e) support roll planetary rolls

cage

(f )

Figure 7.11 Some basic types of roll arrangements: (a) two-high, (b) three-high, (c) four-high, (d) cluster, (e) tandem rolling with three stands, and (f) planetary milling. McGraw-Hill Encyclopedia of Science and Technology, Vol. 11, 8th ed., p. 54. Copyright  1997 by McGraw-Hill.

where R is the radius of the roller and θ is the angle subtended by the roll center by the vertical and the point of interest at the roll surface. H takes on a value of H0 when θ = θ0 , the angle at the plane of entry, which can be approximated by θ0 =



h0 − hf R

(7.8)

The speed at which the metal moves through the rollers must change in order to keep the volume rate of flow constant through the roll gap. Hence, as the thickness decreases, the velocity increases. However, the surface speed of the roller is constant, so there is relative sliding between the roller and the metal. The direction of the relative velocity changes at a point along the contact area called the neutral or no-slip point (point N in Figure 7.10). At the neutral point, the roller and metal have the same

PROCESSING OF METALS AND ALLOYS

697

Cooperative Learning Exercise 7.1

The neutral plane (point N in Figure 7.10) is the point at which the subtended angle is θ = θN , and can be found in several ways. A precise method is to balance the entry and exit forces pressures; i.e., Pe = Px and solve for H at the neutral plane, HN , which can then be substituted into Eq. (7.7) to solve for θN . Alternatively, there are empirical relations for θN . Consider the cold-rolling of an aluminum strip that is rolled from 4 mm to 3.3 mm in thickness with a roller 500 mm in diameter. The coefficient of friction is 0.06. Person 1: Use Eq. (7.5) to (7.8) to determine the subtended angle at the neutral point. Hint: First, calculate θ0 from Eq. (7.8), then calculate HN by equating Eq. (7.5) and (7.6) at H = HN . Then use Eq. (7.7) to find θN . Person 2: Use the following empirical relationship to determine the subtended angle at the neutral point  h0 − hf h0 − hf − . θN = 4R 4f R Then help your partner finish his or her calculations. Compare your answers. How similar or different are they? Answers: θN = 0.0141 radians; θN = 0.0148 radians

velocity. Prior to the neutral point, the metal moves more slowly than the roller; to the right it moves faster (Vf > V0 in Figure 7.10). The direction of frictional forces are thus opposite in these two regions. The net frictional force acting on the metal must be in the direction of exit to enable the rolling operation to take place. It is sometimes necessary to use multiple rolling steps in sequence, interspersed with annealing steps to improve plastic deformation. The largest product in hot rolling is called a bloom. By successive hot- and cold-rolling operations the bloom is reduced to a billet, slab, plate, sheet, strip, and foil, in decreasing order of thickness and size. The initial breakdown of the ingot by rolling changes the course-grained, brittle, and porous structure into a wrought structure with greater ductility and finer grain size. 7.1.2.3 Extrusion. Extrusion involves the pressing of a small piece of metal, called a billet or slug, through a die by the application of force. A billet or slug is differentiated from an ingot primarily by its smaller size. The die in extrusion processing is sometimes called an orifice or nozzle. The forces necessary to push the metal through the orifice are generated by a ram that pushes on the billet, generating both compressive and shear forces. Because no tensile forces are involved, fracture is avoided. When the billet is forced through the orifice by a ram, the process is called direct extrusion (see Figure 7.12). It is also possible to force the die down around the billet to cause the same change in cross-sectional geometry, in a process called indirect extrusion. In either case, the efficiency of the extrusion process is measured by the ratio of work necessary for deformation [Eq. (7.3)] to the actual work performed. Extrusion is primarily used for lower-melting point, nonferrous metals (see Table 7.2); but with the aid of lubricants and improvements in presses, the extrusion of steel has become available. Most carbon steels and stainless steels are extruded at temperatures of about 1200◦ C, in which glass is used as a lubricant. Glass is inexpensive, easy

698

PROCESSING OF MATERIALS

Container

Billet Tube Billet Container

Figure 7.12 Schematic illustration of direct extrusion of a metallic billet. From Z. Jastrzebski, The Nature and Properties of Engineering Materials, 2nd ed. Copyright  1976 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.

Table 7.2

Some Typical Processing Parameters for the Extrusion of Nonferrous Metals

Materials

Minimum Section Thickness, (in.)

Extrusion Temperature (◦ F)

Extrusion Pressure (ksi)

Extrusion Exit Speed (fpm)

Extrusion Ratio ln(A0 /A1 )a

Aluminum alloys Copper alloys Magnesium alloys Zinc alloys

0.04–0.30 0.05–0.30 0.04 0.06

550–1050 1200–1650 570–800 400–660

40–130 30–130 100–130 90–110

4–300 80–1000 4–100 75–100

4–7 5–6.5 4–5.3 4–5

a

A0 is the cross-sectional area of the billet; A1 is the cross-sectional area of the extruded shape. Source: T. Altan, S.-I. Oh, and H. L Gegel, Metal Forming: Fundamentals and Applications. Copyright  1983 by the American Society for Metals.

to use, and chemically stable, and it assists in dissolving oxide layers on the metal surface. The extrusion process is commonly used to form tubes and other continuous shapes of complex cross section (see Figure 7.13). We will describe the extrusion process in more detail in Section 7.3 within the context of polymer processing. 7.1.3

Powder Metallurgy

The process of converting metallic powders into ingots or finished components via compaction and sintering is called powder metallurgy. Powder metallurgy is used whenever porous parts are needed, whenever the parts have intricate shapes, whenever the alloy or mixture of metals cannot be achieved in any other manner, or whenever the metals have very high melting points. The technique is suitable to virtually any metal, however. An organic binder is frequently added to the metallic powders during the mixing state to facilitate compaction, but volatilizes at low temperatures during the sintering process. The advantages of powder metallurgy included improved microstructures and improved production economies. Although the cost of the powders can be greater than that of ingots and mill products obtained by casting or wrought processing, the relative

PROCESSING OF METALS AND ALLOYS

699

Some product forms (not to scale)

Beams

Standard rails

Angles Tees

Crane rails

Round Hexagonal Square Octagonal

Zees Channels

Piling

Joint bars

Flat Half round Triangular

Figure 7.13 Some cross-sectional geometries possible from metal extrusion. Reprinted, by permission, from J. F. Shackelford, Introduction to Materials Science for Engineers, 5th ed., p. 421. Copyright  2000 by Prentice-Hall, Inc.

ease of forming the powder into a final product more than offsets the labor costs and metal losses associated with the other two processes. The powder metallurgy process consists of three basics steps: powder formation; powder compaction; and sintering. Each of the steps in powder metallurgy will be described in more detail. 7.1.3.1 Powder Formation. Metallic powders can be formed by any number of techniques, including the reduction of corresponding oxides and salts, the thermal dissociation of metal compounds, electrolysis, atomization, gas-phase synthesis or decomposition, or mechanical attrition. The atomization method is the one most commonly used, because it can produce powders from alloys as well as from pure metals. In the atomization process, a molten metal is forced through an orifice and the stream is broken up with a jet of water or gas. The molten metal forms droplets to minimize the surface area, which solidify very rapidly. Currently, iron–nickel–molybdenum alloys, stainless steels, tool steels, nickel alloys, titanium alloys, and aluminum alloys, as well as many pure metals, are manufactured by atomization processes. Chemical powder formation methods include the reduction of oxides, reduction of ions in solution by gases, electrolytic reduction of ions in solution, and thermal decomposition of gaseous molecules containing metal atoms. Sponge iron powder, for example, is produced by reacting a mixture of magnetite (Fe3 O4 ) ore, coke and limestone. The product is crushed to control the iron particle size, and the iron is magnetically separated. Iron powder can also be produced by hydrogen reduction of ground mill scale and the reaction of atomized high-carbon steel particles with iron oxide particles. Copper, tungsten and cobalt powders can be produced by hydrogen reduction of oxide powders. Nickel and copper powders can be produced by hydrogen reduction of their corresponding amine sulfates, and iron and nickel powders of high purity and fine particle size are produced by thermal decomposition of their gaseous carbonyls, Fe(CO)5 and Ni(CO)4 , respectively. Mechanical methods of powder production, also known as mechanical attrition, typically require a brittle material, or at least one that becomes brittle during processing. There are various types of equipment that perform particle size reduction,

700

PROCESSING OF MATERIALS

107 Jaw and gyratory crushers

106

Ring roller and bowl mills

Feed size, µm

105 Hammer mills

104 103 102

Ball, pebble and rod mills

101

Stirred media mills

100 105

104

Vibratory mills

Fluid energy mills

103 102 101 Product size, µm

100

10−1

Figure 7.14 Classification of equipment used in mechanical attrition. Reprinted, by permission from ASM Handbook, Vol. 7, Powder Metallurgy (1984), ASM International, Materials Park, OH 44073-0002. p. 70.

termed comminution, but they are generally called mills and operate on the principles of grinding, crushing, and particle fracture (see Figure 7.14). In a device known as a ball mill, spherical milling media are used inside a vial that is agitated via any one of a number of mechanical means to cause ball–ball and ball–vial collisions. The metallic powder is trapped during these collisions. Figure 7.15 shows the process of trapping an incremental volume of metallic powder between two balls. Compaction begins with a powder mass that is characterized by large spaces between particles compared with the particle size. The first stage of compaction starts with the rearrangement and restacking of particles. Particles slide past one another with a minimum of deformation and fracture, producing some irregularly shaped particles. The second stage of compaction involves elastic and plastic deformation of particles. Cold welding may occur between particles in metallic systems during this stage. The third stage of compaction, involving particle fracture, results in further deformation and/or fragmentation of the particles. In the case of metals, fracture is the result of work hardening and embrittlement (cf. Section 5.1.2.2). For brittle materials, particle fracture is well described by Griffith theory (cf. Section 5.2.1.2). Cooling of the metal below the brittle–ductile transition, such as with liquid nitrogen in cryogenic milling, can aid in increasing brittle fracture during milling. As fragments decrease in size, the tendency to aggregate increases, and fracture resistance increases. Particle fineness approaches a limit as milling continues and maximum energy is expended. The major factors contributing to grind limit are increasing resistance to fracture, increasing cohesion between particles, excessive clearance between impacting surfaces, coating of the grinding media by fine particles, surface roughness of the grinding media, bridging of large particles to protect fine particles, and increasing apparent viscosity as the particle size decreases. We will see that these milling methods are also useful for the formation of powders and particles from brittle materials such as ceramics.

PROCESSING OF METALS AND ALLOYS

(a)

(c )

701

(b )

(d )

(e )

Figure 7.15 Schematic illustration of compaction and fracture processes in the formation of metallic powders via ball milling. ASM Handbook, Vol. 7, Powder Metallurgy (1984), ASM International, Materials Park, OH 44073-0002. p. 2.

7.1.3.2 Compaction and Hot Pressing. The shaping of components from metallic powders is accomplished by compacting the particles under pressure, and sometimes at elevated temperatures. Typically, compaction temperatures are in excess of 75% of the absolute melting temperature, and compaction times are on the order of 103 to 104 s. The compaction operation is performed primarily to provide form, structure, and strength, known as green strength, necessary to allow for handling before sintering. Compaction at room temperature involves pressures of 70–700 MPa, depending on the powder, and green densities (density prior to sintering) of 65–95% of theoretical (powder material) densities are achievable. Pressures for hot pressing are usually lower due to die or pressure chamber limitations. The density–pressure relationship for powder compaction at room temperature typically increases from the apparent density at zero pressure to values that approach the theoretical density at high pressures, as illustrated in Figure 7.16. A compact with 100% theoretical density would indicate that it contains no porosity. Soft powders are more easily densified than hard powders at a given pressure, and irregularly shaped powders have lower densities than spherical powders in the low-pressure regime. The mechanism of densification begins with individual particle motion at low pressure before interparticle bonding becomes extensive. At higher pressures, the compact

702

PROCESSING OF MATERIALS

theoretical density

Density

apparent density

Pressure

Figure 7.16 Pressure-density relationship in the compaction of metallic powders. Reprinted, by permission, from Encyclopedia of Materials Science & Engineering, Vol. 5, p. 3874. Copyright  1986 by Pergamon Press.

acts as a coherent mass and densifies with decreased interparticle motion. At pressures of about 100–700 MPa, it has been shown that the slope of the density–pressure curve is generally proportional to the pore fraction, 1 − ρr : dρr ≈ K(1 − ρr ) dP

(7.9)

where ρr is the ratio of actual compact density to theoretical density (ρ/ρt ), P is the pressure, and K is a proportionality constant that increases as the powder becomes softer. In general, K is proportional to the reciprocal of the nominal yield strength of the powder material. Powders can be compacted at room temperature using a so-called die and punch design, in which the powders are compacted in a tool steel or cemented carbide die. The process can be automated, as illustrated in Figure 7.17, so that the die can be filled with powder, the upper and lower portions of the die, called punches, are brought together to compact the powder, and the compact is then ejected to prepare for the next cycle. The powder capacity of these presses can be as high as 1000 kg, and the maximum part size is limited to about 300 cm2 in cross section. As the height-to-width ratio of the die and compact increases, it becomes more difficult to densify the powder in a uniform manner. Thus, the top of the compact may reach a higher density than the bottom. Die compaction of simple shapes can be carried out at elevated temperature using carbide, superalloy, refractory metal, or graphite dies. An inert gas atmosphere or vacuum is often used to protect the die and/or the powder. For example, beryllium powder is compacted at about 1350◦ C in a graphite die under vacuum with pressures of 2–4 MPa. The die-and-punch method of compact imparts an anisotropic pressure to the compact. It is possible to apply pressure equally in all directions using a process called isostatic pressing, in which the powder is compacted with a fluid, such as hydraulic fluid, water, or an inert gas. Isostatic pressing is carried out using formed, flexible molds, called cans, that contain the powder and allow the pressure from the fluid to be transferred to the powder. The flexible molds are then stripped off the compact after densification. Isostatic pressing eliminates the die-wall friction inherent in die compaction and allows for consolidation of large components. However, isostatic pressing

PROCESSING OF METALS AND ALLOYS

703

Upper punch

Compact Die

Lower punch

Die

(a)

Upper punch Die

Compact

Die

Lower punch (b)

Upper punch Compact Die

Lower punch

Die

(c)

Figure 7.17 Schematic illustration of automated die-and-punch process in which (a) die is filled, (b) die is compacted, and (c) compact is ejected. Reprinted, by permission, from Encyclopedia of Materials Science & Engineering, Vol. 11, p. 3875. Copyright  1986 by Pergamon Press.

cannot provide the dimensional tolerances and cycle rates of die compaction. Isostatic pressing is useful for forming intricate shapes and for forming parts that are near their final, desired dimensions prior to sintering, in what is termed near net-shape processing. When the powder is isostatically compacted at elevated temperatures, the process is called hot isostatic pressing (HIP). In this case, the flexible dies are often made of thin metals, and high-pressure gases such as argon are used to heat the part rapidly and reduce thermal losses. Pressure up to 100 MPa and temperatures in excess of 2000◦ C are possible using HIP, and parts up to 600 kg can be fabricated. A schematic diagram of a typical HIP apparatus is shown in Figure 7.18. Metals that are processed commercially by HIP include various specialty steels, superalloys, hard metals, refractory alloys, and beryllium. We will see in Section 7.2 that HIP is also particularly useful for the densification of ceramic components. 7.1.3.3 Sintering and Densification. Recall from Section 2.2.2.4 that sintering is the consolidation of particles during heat treatment, and that it is controlled by the reduction in free energy of the system that accompanies surface area elimination during densification. Since the thermodynamics of sintering have already been described, we

704

PROCESSING OF MATERIALS

Pressure vessel

Gas storage

Coolant inlet

Exhaust valve

Relief valve

Check valve

Pressure gauge

Pressure regulator

Filter

Compressor

Heater Coolant outlet Pressure gauge Gas supply Overpressure release

Thermocouple

Furnace control supply Control circuit Temperature control

Figure 7.18 Schematic diagram of a hot isostatic pressing (HIP) operation. Reprinted, by permission, from Encyclopedia of Materials Science & Engineering, Vol. 3, p. 2188. Copyright  1986 by Pergamon Press.

will concentrate here on the practical aspects of the sintering of metallic particles in powder metallurgy. A similar approach will be used in an upcoming section on the sintering of ceramic powders. For sintering to take place, the temperature must generally be maintained above onehalf the absolute melting point of the material. This condition results in atomic diffusion and neck formation between the powder particles in the solid state (cf. Figure 2.31). In the initial stages of sintering, interparticle bonds formed during compaction grow in size, resulting in a significant increase in strength. The pore channels in the compact then begin to round and become discontinuous. Further thermal processing forms rounded pores that can undergo coarsening. Typically, the ductility of the metal compacts increases as the pore morphology changes. Sometimes, density can decrease during sintering because of gas expansion or interdiffusion phenomena called Kirkendall porosity formation. Various batch and continuous furnaces are employed for sintering operations. Composition and impurity control and oxidation protection are provided by the use of vacuum conditions or inert gas atmospheres. It is important to control the atmosphere to prevent undesired chemical reactions, such as oxidation or carburization. However, some reactions may be desirable during sintering, such as the control of carbon content in steels and oxide removal using a reducing furnace atmosphere. Hydrogen, either pure or diluted with nitrogen, and dissociated ammonia (75%H2 –25%N2 ) are commonly used as reducing atmospheres. Exothermic gas formed by reacting air and natural gas mixtures (10:1 to 6:1) is inexpensive and reduces many materials. Endothermic gas formed by reacting air–natural gas mixtures (2.5:1 and 4:1) is both highly reducing and carburizing and is often used for sintering steel compacts. Additional considerations are involved in the sintering of ceramic components, and these are addressed in Section 7.2. 7.2

PROCESSING OF CERAMICS AND GLASSES

There are perhaps a wider variety of techniques that can be used to form and shape ceramics and glasses than for any of the other materials classes. Much of the variation

PROCESSING OF CERAMICS AND GLASSES

705

in processing is due to the end use of the product. Certain techniques are used for structural ceramics, while others are used for semiconductor, superconductors, and magnetic materials. In general, there are three types of processes we will consider: powder-forming, melt processing, and chemical processing. The melt processing of ceramics and glasses is similar enough to the melt processing of polymers that its description will be delayed to Section 7.3. The latter category, which includes vapor phase and sol-gel processing, is considered optional in this chapter. Two topics related to powder processing, firing and sintering, will be described here. The powder-forming processes are similar in many ways to those used for powder metallurgy described in the previous section. For example, pressing is a common method for processing ceramics; however, ceramic powders can be pressed in either dry or wet form. In wet form, they can also be extruded, just like metals, and cast in a variety of process variations. The nominal forming pressures and shear rates associated with some of these processing methods are summarized in Table 7.3. You may want to refer back to this table when each of the various processing techniques is described in more detail. Before a description of powder processing is initiated, there is a consideration in ceramic-forming that is more prominent than in metal-forming which must be addressed: dimensional tolerance. Post-forming shrinkage is much higher in ceramics processing than in metals and polymer processing because of the large differential between the final density and the as-formed density. Volume shrinkage may occur during drying (Svd ), during the removal of organic binders or reaction of bond phases producing consolidation (Svb ), and during sintering (Svs ). The total volume shrinkage Svtot is then the sum of these contributions: Svtot = Svd + Svb + Svs

(7.10)

Table 7.3 Nominal Processing Pressures and Shear Rates in Some Ceramic Forming Processes

Forming Process

Pressure (MPa)

Shear Rate (1/s)

Pressing Roll/isostatic Metal die Plastic forming Injection molding Extrusion

Varies